Classification jobs

...two fixed cameras (mounted on one rig, overlapping field of view) into a single panoramic image (~180°), using calibration/homography. The cameras don't move relative to each other, so this should be a one-time calibration applied per frame. Motion-density tracking: From the panoramic feed, detect where players are concentrated (background subtraction / foreground blob density, not per-object classification) and use that to drive an automatic pan/crop — i.e. a virtual camera that follows the action without a human operator. Output: a single, smooth, watchable video stream (no visible seam/ghosting at the stitch line, no jittery panning). What this is NOT: we are not looking for a deep-learning/ball-detection/player-identification system. Classical CV techniques (h...

...two fixed cameras (mounted on one rig, overlapping field of view) into a single panoramic image (~180°), using calibration/homography. The cameras don't move relative to each other, so this should be a one-time calibration applied per frame. Motion-density tracking: From the panoramic feed, detect where players are concentrated (background subtraction / foreground blob density, not per-object classification) and use that to drive an automatic pan/crop — i.e. a virtual camera that follows the action without a human operator. Output: a single, smooth, watchable video stream (no visible seam/ghosting at the stitch line, no jittery panning). What this is NOT: we are not looking for a deep-learning/ball-detection/player-identification system. Classical CV techniques (h...

...two fixed cameras (mounted on one rig, overlapping field of view) into a single panoramic image (~180°), using calibration/homography. The cameras don't move relative to each other, so this should be a one-time calibration applied per frame. Motion-density tracking: From the panoramic feed, detect where players are concentrated (background subtraction / foreground blob density, not per-object classification) and use that to drive an automatic pan/crop — i.e. a virtual camera that follows the action without a human operator. Output: a single, smooth, watchable video stream (no visible seam/ghosting at the stitch line, no jittery panning). What this is NOT: we are not looking for a deep-learning/ball-detection/player-identification system. Classical CV techniques (h...

...two fixed cameras (mounted on one rig, overlapping field of view) into a single panoramic image (~180°), using calibration/homography. The cameras don't move relative to each other, so this should be a one-time calibration applied per frame. Motion-density tracking: From the panoramic feed, detect where players are concentrated (background subtraction / foreground blob density, not per-object classification) and use that to drive an automatic pan/crop — i.e. a virtual camera that follows the action without a human operator. Output: a single, smooth, watchable video stream (no visible seam/ghosting at the stitch line, no jittery panning). What this is NOT: we are not looking for a deep-learning/ball-detection/player-identification system. Classical CV techniques (h...

...membership payments using various payment methods. ​Geolocation and Mapping System ​Location Filter: Search based on the client's current address or the location where the work is required via GPS. ​Geographical Visualization: Interactive map to view the concentration of available technicians by state, city, and sector. ​Service Catalog and Classification ​Main Categories: General listing of industries (e.g., Electrical, Refrigeration). ​Classification and Subcategories: Specific breakdown to narrow down the search for the exact service required. ​Advanced Search Filters: Sorting by proximity, price, availability, and reputation. ​Reputation and Rating System ​Performance Metrics: Automatic count of successfully completed services. ​Ratings and Reviews: Comment system...

I have a clean, ready-to-use Iris dataset and need a complete Support Vector Machine classification pipeline built around it. The job is straightforward: tune an SVM (scikit-learn or similar) to separate the three Iris species, document the process, and hand over reproducible code and results. Here’s what I expect: • A concise notebook or script that loads the data, performs any minimal preprocessing you find beneficial, runs hyper-parameter optimisation, trains the final model, and outputs accuracy, precision, recall and the confusion matrix. • A short write-up (markdown inside the notebook is fine) explaining why the chosen kernel and parameters work best, plus any insights you notice in the feature space. • Saved model file so I can deploy or reload it...

This is primarily an Android + backend booking system pro...notifications and teacher responses. The backend/database should control booking status, availability, conflicts, history, and audit records. For example: PENDING → ACCEPTED or PENDING → REJECTED and for modifications: ACCEPTED → MODIFIED with something like: Reason: Teacher unavailable at the originally selected time. This makes the system much more reliable and auditable. Freelancer project classification I would categorize it as: Android App Development + Full-Stack Development + Booking System + Admin Panel + Email Automation It is a medium-complexity full-stack project, even though the Android app itself can be relatively simple. The backend booking logic is the part that needs the most care...

Technology Lead – AI SaaS & Revenue Automation We are building a market-native, autonomous AI revenue engine for European B2B companies. The platform transforms a company’s website into a credible, approved outbound campaign within 30 minutes—then supports prospect verification, controlled outreach, reply classification and appointment booking without requiring a lengthy CRM implementation. We are seeking an experienced Technology Lead to take ownership of the technical architecture and oversee a remotely distributed team of software engineers. This is a hands-on leadership role for someone who can translate product strategy into reliable software, establish engineering standards and ensure consistent delivery across the team. Key responsibilities • Own th...

...contributed to the model's prediction. Built REST APIs using FastAPI for image upload, prediction, model information, and health monitoring. The platform also supports prediction history and is designed for integration with AI-assisted medical report generation using LLMs. Key Features 1. Medical Image Analysis – Processes chest X-ray images using deep learning. 2. Disease Classification – DenseNet121-based classification for 5 medical categories. 3. Explainable AI – Grad-CAM generates heatmaps to visualize important image regions. 4. FastAPI REST API – Provides scalable endpoints for image prediction and model services. 5. Prediction Results – Returns predicted condition and confidence score. 6. AI-Assisted Reporting – Designed to...

...broad contours—coverage must span almost all languages relevant to the regions under study, while drawing from mainstream news outlets, official government releases and authoritative expert commentary. What I need now is a small, cross-disciplinary group comfortable with media monitoring, foreign-policy analysis and-or modern AI / machine-learning techniques (NLP, topic modelling, LLMs, classification workflows in Python or R). Together we will review and refine: • the critical sources in each language and define ingestion methods • Build or adapt an AI layer that flags emerging narratives and policy signals for human review • an oversight framework that lets analysts verify, contextualise and, where necessary, override the algorithm’s out...

...DeepForest or other tree detection/segmentation and species-classification frameworks. The expected workflow may include: Remote Sensing Imagery → Individual Tree/Crown Detection → Feature Extraction → Species Classification → GIS Species Map & Tree Count Potential data sources may include: * High-resolution satellite imagery * Multispectral imagery * Sentinel-2 time-series data * Drone imagery, where required * Ground-truth/GPS data from field surveys The model should ideally make use of **spectral, temporal, textural and/or crown structural features** where appropriate. Expected Deliverables * Working AI/ML model and complete source code * Pre-processing and training pipeline * Individual tree/crown detection or segmentation * Classifica...

...fiscal years 2019 through 2025. The universe should be drawn exclusively from Refinitiv Eikon or Bloomberg terminals and limited to three sectors: Technology, Healthcare, and Consumer Goods. The core of the job is to pull three financial performance indicators for every firm–Revenue, Net Income, and EBITDA–together with essential identifiers (company name, ISIN, ticker, country, and sector classification). Because the focus is strictly “non-financial,” please exclude banks, insurance companies, and other financial services entities. Preferred format is a clean, well-labeled Excel or CSV file that I can import straight into Stata/R. Please include column headers, year stamps, and your original source codes so the figures can be easily cross-verified. D...

Hello, I have been working in data annotation for almost 3 years, gaining extensive experience in annotations. This makes me a valuable addition to your team. In addition, I have much experience in image annotation, segmentation, bound...image annotation, segmentation, bounding boxes, polygons, key points, 2D and 3D annotations, and even LIDAR annotations. Tools: CVAT Roboflow LabelImg Labelbox VGG Doccano Label Studio Annotation Solutions: Bounding Boxes Image annotation Object labeling/tagging Semantic Segmentation Polygons Annotation/masks Polylines Annotation Key Points annotation Sentiment, Text & Topic Analysis Image classification and categorization Object Tracking Data Ann...

...Learning Areas The training should cover: **1. AI Fundamentals** * Artificial Intelligence concepts and terminology * How modern AI systems work * Real-world AI use cases * AI models, data, training and inference * Responsible AI, governance and AI ethics **2. Machine Learning Fundamentals** * Machine Learning concepts explained from the basics * Supervised and unsupervised learning * Classification, regression and clustering * Training, testing and evaluating models * Basic Python required for AI/ML * Practical ML exercises using real datasets **3. Generative AI** * Generative AI and Large Language Models (LLMs) * Prompt engineering * Embeddings and vector databases * Retrieval-Augmented Generation (RAG) * AI agents and agentic workflows * Working with AI APIs * Building ...

... DNC status must prevent future automated contact. Store the source, reason and timestamp. Ordinary users cannot bypass it. Unauthorized overrides must be rejected and logged. The architecture should support future SMS STOP handling. ## 13. Usage and billing ledger Track each call’s organization, agent, duration, Telnyx cost, TelesalesOS charge, recording/transcription costs and billing classification. Business rules: * Agents receiving free company leads do not personally pay telco charges. * Other agents may pay their own usage. Create an auditable ledger that can later connect to Stripe. Use integer minor units or precise decimal fields, never floating-point money calculations. ## 14. Webhooks and reliability Telnyx webhooks must be: * Signature verified * Idemp...

...learning classification models to enhance anomalous behavior detection. • Achieved 88% accuracy in identifying high-risk taxpayers through advanced predictive modeling. • Developed a scalable analytics pipeline using PySpark for distributed data processing and an interactive Streamlit dashboard with 7 analytical views with KPIs, Tables and Graphs for real-time compliance monitoring. Built an end-to-end ETL data analytics pipeline with 500K+ transaction records for customer purchasing insights. • Preprocessed the dataset, performed an 80/20 train-test split, and conducted EDA and feature engineering. • Designed customer segmentation ML models using RFM analysis to classify users into 3 behavioral groups. • Improved decision-making and strategic planning by...

...Learning Areas The training should cover: **1. AI Fundamentals** * Artificial Intelligence concepts and terminology * How modern AI systems work * Real-world AI use cases * AI models, data, training and inference * Responsible AI, governance and AI ethics **2. Machine Learning Fundamentals** * Machine Learning concepts explained from the basics * Supervised and unsupervised learning * Classification, regression and clustering * Training, testing and evaluating models * Basic Python required for AI/ML * Practical ML exercises using real datasets **3. Generative AI** * Generative AI and Large Language Models (LLMs) * Prompt engineering * Embeddings and vector databases * Retrieval-Augmented Generation (RAG) * AI agents and agentic workflows * Working with AI APIs * Building ...

Please Sign Up or Login to see details.

I'm looking to create engaging skits using computer vision technology for entertainment purposes. Key Requirements: - Develop creative and entertaining skit concepts leveraging computer vision. - Implement computer vision techniques like object detection or image classification. - Collaborate on scripting and staging for the skits. Ideal Skills: - Strong background in computer vision. - Creative thinking and experience in entertainment production. - Good scripting and storytelling skills.

...chemical data contained in our SDS Reference Cards.** This is NOT simple data entry. ## YOUR TASK You will review SDS cards and verify whether the chemical information is scientifically correct and consistent. Depending on the compound, you will check: * CAS number * chemical identity * chemical name / synonyms * molecular formula * molecular weight * structure / SMILES / InChI * GHS / CLP classification * hazard statements * precautionary statements * toxicological information * ecological information * physical and chemical properties * transport information * regulatory information * source references * consistency between different sections of the SDS You will compare the information against authoritative scientific and regulatory sources where appropriate. ## THE OBJE...

...low-competition commercial queries We are especially interested in **long-tail keywords with commercial intent and low competition**. We do NOT want generic keywords such as simply: "chemical" "chemicals" "chemical supplier" unless supported by a clear commercial strategy. ## DELIVERABLE Create a complete Chinese SEO structure for MolGod including: 1. keyword research; 2. search-intent classification; 3. competition analysis; 4. keyword prioritization; 5. proposed URL structure; 6. Chinese landing-page architecture; 7. internal linking structure; 8. title/H1/meta recommendations; 9. schema/structured-data recommendations; 10. recommendations for CAS-specific SEO pages; 11. recommendations for REACH/SDS/CLP/export pages; 12. recommendations f...

...Business observations Physical location found: Yes/No Business signage visible: Yes/No Premises open: Yes/No Staff observed: Approximate number Customers observed: Approximate number Customer activity: Low / Moderate / High Services/products visibly being provided Any visible indication of active transactions General condition of premises Any unusual circumstances 8. Operational Status Classification The final report should use one of these conclusions: Clearly Trading Normal customer/staff activity and/or visible business transactions/services were observed. Appears to Be Trading Some indications of ongoing operations were observed, but activity was limited during the visit. Unable to Confirm The premises was found, but insufficient activity was observed to confide...

... wiring, testing, and troubleshooting for components like three-phase induction motors, VFDs, ESP32-based data acquisition systems, harmonic filters, capacitor banks, and more. 2. Simulation with ETAP: Load Flow Analysis, Harmonic Analysis/THD, Voltage Unbalance scenarios, and comparison between simulation results and physical prototype measurements. 3. Data Analysis/Intelligent Diagnosis: Classification of operating conditions (e.g., normal operation vs disturbances) and implementation of corrective actions with performance evaluation. Ideal Skills & Experience: - Expertise in Power Quality | Induction Motors | VFDs | ETAP | Harmonic Analysis | Embedded Systems | Hardware Prototyping - Practical experience with laboratory/industrial hardware is essential. Deliverables: - T...

I’m building a browser-accessible tool that can take an image uploaded by the user, run it through an AI model, and immediately return classification or detection results on-screen. All core logic must live server-side, exposed through a clean REST or GraphQL endpoint, so the front-end remains lightweight and responsive across modern web browsers. Key expectations • Model accuracy matters: please start with a proven open-source architecture (e.g., YOLOv8, ResNet, EfficientDet) fine-tuned on a small sample set I’ll provide, then document how to retrain it when new data arrives. • One-click deploy: include a Dockerfile and concise README so I can spin everything up on a fresh VPS. • Results returned as JSON plus visual overlays (bounding boxes or mask...

...features can be added without requiring a complete redesign. Our Services Compass assists clients and families with subjects including preparation before reporting, reception and intake, diagnostics and assessment, classification, facility assignment, daily institutional life, family communication, visitation, commissary and financial accounts, educational/program opportunities, and reentry planning. Again, our role is educational and consultative. We do not represent clients legally, provide legal advice, or claim the ability to influence correctional placement, classification, release, or other government decisions. Ideal Freelancer I am looking for someone who: Has strong experience building professional Wix websites Has a portfolio demonstrating polished business/ser...

...and professionals organize their accounting and financial records so they can focus on running their business. I am available for projects involving: • Bookkeeping and accounting • GST and taxation support • Income tax-related accounting work • Bank and ledger reconciliation • Financial statement preparation • Audit assistance and documentation • Accounts payable and receivable • Expense classification and ledger management • Accounting data entry and cleanup • Financial reporting • Reviewing accounts and identifying discrepancies • Ongoing monthly accounting support Whether you need your books cleaned up, accounts maintained regularly, transactions properly classified, or assistance with GST, taxation, and audit-rela...

...Deliverables The final deliverables should include: A written report describing the search methodology and findings. Quantitative results and categorization of the references identified. A comprehensive Excel or similar source log identifying each relevant result, source/platform, URL, date located, date of content where available, search query through which it was identified, and product-category classification. Supporting links, screenshots or other source documentation where appropriate. Qualifications Candidates should have: A Master’s degree or doctorate preferred in computer science, information science, data science, information retrieval, search technology, Internet/web analytics, digital forensics, or a closely related technical field. A substantial academic an...

I need an application to assess department performance based on various faculty achievement KPIs. Key KPIs to measure: - Research publications - Teaching effectiveness - Community engagement - Book publications - Patencies - Professional classification Ideal skills and experience: - App development - Experience with performance metrics - Data visualization - User-friendly interface design Please include relevant past work in your application.

...APIs, database integration, and backend services. Built using Node.js, , MongoDB, React.js, and Python/NLP. AutoPulse – Used Car Review Sentiment Analysis () Built an NLP-based system for analyzing customer reviews from used-car platforms. Collected and processed large-scale review data. Performed text preprocessing, cleaning, and sentiment classification. Analyzed customer opinions around aspects such as price, car condition, delivery, documentation, mileage, and customer service. Worked with machine learning/deep learning approaches for sentiment analysis. Used Python, Pandas, NLP, Matplotlib, and ML/DL techniques. Online Compiler() Developed an online coding platform where

...same-day Tally Prime data entry for my small business so every figure in our books mirrors the reality of our bank and cash positions. The work centres on three concrete areas: • Ledger management – creating, updating and aligning all ledgers with our chart of accounts. • Journal, payment, receipt and contra entries – posting each transaction with the right narration and tax codes. • Expense classification and fresh ledger creation – grouping costs logically and opening any new heads that surface. Please post everything directly inside our existing Tally Prime data file, observe GST and TDS rules already configured and keep audit trails intact. A spot check of random vouchers must reconcile 100 % with source documents before hand-off; that i...

I’m putting together a fresh lead-generation list of businesses in the U...those sites, confirm the industry sector, and capture the decision-maker’s best contact details. I need, at minimum, the main email address and a working phone number for each domain you collect. Please deliver the data in an Excel spreadsheet with clear column headings so I can drop it straight into my CRM. Acceptance criteria – the file must include: • Business name and URL • Industry classification (retail, hospitality, or healthcare) • Footer phrase found • Contact email • Contact phone • Date data confirmed Clean, deduplicated records only, and each line should reference a live site I can click and verify. rate is $50 per 1000 records Provide...

... * Records of Processing Activities (RoPA). * Privacy Impact Assessments (PIA) and Data Protection Impact Assessments (DPIA). * Data Transfer Impact Assessments (DTIA). * Third-Party Risk Assessments (TPRA). * Data Processing Agreements (DPA). * Data retention, deletion, anonymization, and disposal. * Personal data breach assessment and notification. * Privacy by Design and Default. * Data classification and handling requirements. * Employee privacy and HR processing. * Vendor and outsourcing compliance. * Audit evidence and regulatory readiness. * NDMO Data Management and Personal Data Protection Standards. * Interaction between privacy requirements and Cybersecurity, Risk, Compliance, Legal, and Data Governance controls. International privacy frameworks such as GDPR, ISO 27701...

...Data Extraction & Structuring Convert unstructured web/document content into structured datasets. Design and maintain data schemas. Use rules, regex, layout-aware extraction, NLP/NER, and LLM-assisted extraction where appropriate. Normalize dates, amounts, units, names, addresses, identifiers, and other inconsistent values. Implement deduplication and entity resolution. Maintain taxonomies, classification logic, and reference/master data. 4. Data Quality & Validation Build automated quality checks for completeness, accuracy, freshness, consistency, uniqueness, and validity. Implement validation and anomaly-detection mechanisms. Maintain gold-standard datasets and benchmark extraction/OCR accuracy. Create human-in-the-loop review processes for low-confidence results. Re...

Freelancer job title U.S. Licensed Customs Broker Needed — HTS Classification & Pre-Import Compliance Review for China-to-USA Retail Kit Job advertisement We are a U.S. New Mexico LLC preparing our first commercial import from Qingdao, China, for sale through Amazon FBA in the United States. We are seeking an active U.S. Customs and Border Protection licensed customs broker—not a general freight forwarder, sourcing agent, or unlicensed consultant—to provide an independent pre-import customs-compliance assessment and written HTSUS classification review. The first shipment is complete and ready to ship. We plan to import the same product repeatedly, so we want to establish a correct, repeatable compliance process before moving the shipment and before pl...

...potentially cardiac surgery, and examining dimensions such as: * Publication trends over time * Evolution of AI/ML methods * Clinical applications and data modalities * Internal vs external validation * Prospective evaluation * Real-world clinical implementation * Explainability and reporting quality I am looking for someone who can help refine the research question, develop the search strategy and classification framework, extract and clean bibliographic data, perform the statistical/bibliometric analysis, and contribute to a publication-ready manuscript. The final study should be methodologically strong and sufficiently distinct from the reference study rather than simply replicating it in another specialty. Please share your previous experience with bibliometric or medical...

Development of a secure, scalable, automation-first, multi-tenant compliance platform for Corporate Clients, Outsourced DPOs, and DPCO Firms. Scope includes Laravel REST API + React/, Mysql, role-based portals, strict tenant isolation, automated regulatory classification, RoPA/compliance automation, evidence vault, DSAR and breach workflows, active cookie consent controls, subscriptions/payments, audit/reporting, Nigerian production deployment, documentation, training, and 60-day post-launch support. Fixed Price: USD $1,100 Timeline: 9 Weeks Payment: Milestone-based via Freelancer.com Escrow.

I’m ready to move fast on an AI-powered application whose sole focus is data analysis for my business clients. The core dataset you’ll be working with includes detailed information about business owners—everything from demographic fields to transaction histories and sector classification. Additionally, I have Excel data for specific business sectors like General Practice, Pharmacy, and Child Care, containing details such as practice name, location, address, contact number, email address, and owner(s). Here’s what I need the model to surface: • Clear purchase-pattern trends over time • Actionable customer-preference clusters • Early churn-risk signals • Identification of owners in a specific sector who may be inclined to sell the...

I offer various options to help you organize and clean up your accounting records in QuickBooks. Here are the services I provide: - Review and Correction of Transactions: I examine existing transactions in QuickBooks Online and correct any errors or discrepancies in the accounts. - Bank Reconciliation: I ensure that bank transactions match the records in QuickBooks Online. - Classification of Expenses and Income: I correctly organize and classify transactions into the appropriate categories. - Duplicate Removal: I look for and eliminate duplicate transactions or unnecessary records that may inflate financial reports. - Inventory Regularization: If you manage inventory, I verify the accuracy of inventory records and adjust quantities as necessary. - Missing Data Update: I complete an...

...our onboarding and off-boarding data perfectly aligned across our media software and the location classification tool. Your core job is straightforward: capture each new retail list I supply, enter it into the media platform with zero typos, flag the corresponding removals, and circulate the updated lists to the internal planning team and our external media partners. Accuracy in every cell is non-negotiable; a single digit out of place pushes ads to the wrong store. While speed and clear communication certainly help, I measure success first by pristine data. Day to day you will log into the media software, upload or remove locations, tag each record correctly using the location classification tool, audit the output, and confirm that the final IDs and tags match before I ...

I have a structured numerical dataset ready for a clean, reproducible classification workflow. The goal is to build, tune, and evaluate two models—Support Vector Machine and logistic regression—then present the results in a way that lets me decide which approach to take to production. Here’s the flow I have in mind: • Pre-process and explore the data (handle missing values, scale where needed, visualise key relationships). • Implement both classifiers in Python with scikit-learn, using cross-validation and grid/random search for hyper-parameter tuning. • Produce clear metrics (accuracy, precision-recall, ROC-AUC) and concise plots that compare the two models side-by-side. • Package everything in a well-commented Jupyter notebook plus a sh...

...hand the SEO workload to a specialist who can make that happen. The focus begins with solid keyword research that will underpin every optimisation move we take next. Scope • Conduct deep keyword research specifically aimed at traffic growth, covering all three engines I rely on—Google, Bing and Yahoo. • Deliver an organised keyword list that includes search volume, difficulty, intent classification and quick-win opportunities. • Provide a concise strategic brief explaining how these terms can be integrated into current pages and future content. Acceptance criteria 1. At least 150 actionable keywords, segmented by intent (informational, commercial, transactional). 2. Data source for search volume and difficulty clearly cited for each term. 3. R...

I’m building an AI-powered workout coach that runs smoothly both in a mobile app and a browser. The core of the p...workout calendar. • Each generated workout includes exercise name, reps/sets (for strength) or hold times/flows (for yoga) plus short video or GIF demos. • Adjustments after a completed session are reflected in the next recommended workout without manual refresh. • UI passes standard accessibility checks and loads in under two seconds on 4G. If you have existing modules for exercise classification or progressive load algorithms, feel free to suggest them. Let me know which frameworks you prefer for the mobile wrapper (React Native, Flutter, etc.) and how you plan to keep the web client in sync. I’m ready to get started as soon as we lo...

I’m putting together an NLP-driven product and need a hands-on developer who works comfortably in Python and prefers PyTorch as the framework of choice. The immediate goal is to design, train, and fine-tune language models that will power features such as text classification, entity extraction, and semantic search. You should be able to: • architect and code end-to-end pipelines—data ingestion, preprocessing, model training, evaluation, and deployment; • experiment with state-of-the-art transformer architectures, optimise hyperparameters, and benchmark performance on agreed metrics; • containerise and push models to a cloud endpoint (AWS, GCP, or Azure—choose what you know best); • document the approach clearly so the team can reproduce...

...three capabilities: reliable text classification, accurate sentiment analysis, and a conversational chatbot that can surface answers from our own content. My training corpus is ready: a large stream of social-media posts plus SEO-focused text collected from our marketing team. I can supply it in JSON or CSV along with basic labels where they already exist. You’ll need to clean, augment, and align this data so it fits neatly into Claude’s fine-tuning workflow (or an equivalent retrieval-augmented approach if that proves more efficient). Here’s what I expect as concrete deliverables: • An end-to-end Claude-compatible training pipeline (data prep, experiment tracking, version control). • Three fine-tuned models or prompt templates covering cl...

We have a trained, working YOLO-based object detection + classification model (Python) that detects and grades apples in tray images as Grade A / Grade B • Need it wrapped into a callable REST API (FastAPI or Flask preferred) that accepts an image and returns grading results • Must be optimized for GPU inference (RunPod, RTX A5000) — not CPU-only • Model loading must be efficient (loaded once, kept warm — not reloaded per request) • Needs to handle real-world phone-camera photos, not just clean training images — proper image preprocessing/normalization required • Robust error handling: bad images, no apples detected, corrupted uploads, timeouts • Must support multiple concurrent requests without performance degradation (10+ concu...

I need a compact framework that lets me evaluate a data-driven model, benchmark it against reasonable baselines, and then roll those findings into a lightweight “sudo” (pseudo) prediction routine I can run or extend on my own. The job breaks down into three clear pieces: 1. Design an evaluation pipeline that captures the usual classification/regression metrics and can be adapted to new datasets with minimal code changes. 2. Wire in a benchmarking step so I can see how alternative algorithms or configurations stack up side-by-side—speed and accuracy both matter. 3. Deliver a working prediction script or notebook that reproduces the best-performing setup from the benchmark and outputs predictions in a clean, documented format. I’m comfortable with e...

...will start with a clean CSV export that already includes labelled outcomes. Feel free to work in Python with scikit-learn, XGBoost, LightGBM, or a comparable library—whatever gets the best accuracy while keeping inference times low. I’m open to simple baseline models first, followed by feature engineering and hyper-parameter tuning to squeeze out extra performance. Because this is strictly a classification task, success is measured by precision, recall, F1 and a well-calibrated ROC-AUC on a hold-out test set. I’d also like a brief explanation notebook so non-technical stakeholders can understand how key features influence the prediction. Deliverables • Clean, commented source code and environment file • Trained model artefact (pickled or equivalent...

I have a collection of unstructured data—mixed text documents and images—and I need an unsupervised learning workflow that reliably flags unusual or suspicious examples. The aim is purely anomaly detection; no labels are available and none can be added, so clustering or classification won’t help here. Here’s the flow I have in mind: • Data preparation: consistent preprocessing for both modalities (tokenisation or embeddings for text, feature extraction for images). • Model development: an unsupervised architecture such as auto-encoder, variational auto-encoder, deep clustering, or another approach you can justify for anomaly detection. • Evaluation: quantitative metrics (reconstruction error distributions, AUC, or similar) plus a concise ...

...completed photo layout should be downloadable as a high-resolution: * JPG/JPEG * PNG Ideally, image quality should be sufficient for: * Clinical documentation * Presentations * Publications * Before-and-after comparisons ## AI / Technical Approach I am open to recommendations regarding the technology. Possible technologies include: * OpenAI Vision API or another vision model for photograph classification * Python * OpenCV * Pillow * React / or similar web frontend * Face/dental landmark detection if helpful **I am not looking for generative AI to modify or recreate the patient's teeth or face.** AI should primarily be used to **recognize, classify, orient, and assist in positioning the original clinical photographs**. ## Privacy / Security Because these are patie...

I have a structured dataset that needs to be turned into a production-ready binary classifier. The records are purely numerical, with no text or image fields involved, and I would like the final solution built around XGBoost. Here is what I need from you: • Clean, explore, and engineer features from the raw numerical data. • Train and fine-tune an XGBoost model for binary classification, validating it with k-fold cross-validation. • Supply well-commented Python code (preferably in a single notebook or script) alongside a brief read-me so I can reproduce your results. • Deliver performance metrics—accuracy, precision, recall, and F1—plus a confusion matrix so I can judge how the model behaves on unseen data. • Package the trained model (...