Horse Racing ML Engineer Needed

@sajjadtaghvaeifr

Hello, I trust you're doing well. I am well experienced in machine learning algorithms, with nearly a decade of hands-on practice. My expertise lies in developing various artificial intelligence algorithms, including the one you require, using Matlab, Python, and similar tools. I hold a doctorate from Tohoku University and have a number of publications in the same subject. My portfolio, which showcases my past work, is available for your review. Your project piqued my interest, and I would be delighted to be part of it. Let's connect to discuss in detail. Warm regards. please check my portfolio link: https://www.freelancer.com/u/sajjadtaghvaeifr

@shadabkhan92

With a deep specialization in machine learning, an extensive Python skill set, and a wealth of experience in building AI models that are not only effective but production-ready, my team and I are the ideal choice for your Horse Racing Intelligence Platform. Having delivered similar end-to-end projects that involved designing, implementing, and retraining ML models using large historical datasets, we are well-versed with the kind of data-driven learning capabilities you are seeking while preserving existing rule-based analytics. To model the win, place, and probability estimation accurately for your horse racing platform, we've extensively utilized Scikit-learn, XGBoost, LightGBM in conjunction with feature engineering using Pandas and Numpy. Given your emphasis on continuous retraining and explainability for predictions, our skill at maintaining ML model reliability without replacing rule-based scoring would prove invaluable for your requirements.

@mykhaylob4

✋ Hi there. I can build ML models for your horse racing platform with feature engineering pipelines, calibrated probability estimates, and integration into your existing FastAPI backend. ✔️ I have rich experience building sports analytics prediction models and recently delivered a similar system using XGBoost with calibrated probabilities and reproducible pipelines. I will develop this by building feature engineering pipelines from race cards, trackwork, and historical results, implementing XGBoost classifiers for win and place with probability calibration using isotonic or Platt scaling , designing a robust handling strategy where missing data reduces confidence rather than creating fabrications , building a model evaluation pipeline with backtesting and ROI analysis, and integrating ML outputs into your existing Intelligence Engine via FastAPI endpoints with model versioning for continuous retraining . I will ensure every prediction remains explainable through feature importance tracking and maintainable code with full documentation . Please click the 'Chat' button to start our valuable conversation. Looking forward to collaborating with you! Best regards, Mykhaylo

@vishal1145

Your ML pipeline will fail in production if your probability calibration doesn't account for class imbalance in horse racing outcomes - where 90% of horses lose and favorites win only 33% of the time. Without proper Platt scaling or isotonic regression, your confidence scores will be systematically overconfident and unusable for betting decisions. Quick question - are you currently handling track condition drift between training and inference? And what's your strategy for cold-start predictions when a horse has limited race history but strong trackwork signals? Here's the architectural approach: - FEATURE ENGINEERING: Build time-decay weighted features from race history, trackwork velocity metrics, and jockey/trainer performance vectors that preserve temporal causality without data leakage. - XGBOOST + LIGHTGBM ENSEMBLE: Implement gradient boosting models with SHAP explainability so every prediction traces back to specific evidence (recent form, track bias, pace figures). - POSTGRESQL INTEGRATION: Design materialized views for feature extraction that refresh incrementally, avoiding full table scans on 500K+ race records during retraining cycles. - FASTAPI ML ENDPOINTS: Expose calibrated probabilities through /predict and /explain routes that return confidence intervals alongside factor attribution, integrating cleanly with your existing Intelligence Engine. - CONTINUOUS RETRAINING: Build Airflow DAGs that retrain models weekly on expanding historical data while monitoring calibration drift using Brier score decomposition. I've built similar probabilistic ranking systems for a DFS sports platform that processed 2M predictions daily with sub-100ms latency. The key was treating this as a learning-to-rank problem rather than pure classification. Let's schedule a 15-minute call to walk through your current schema and discuss whether you need separate models per track surface or a unified approach with surface embeddings.

@muhammadali33345

Build data-driven learning capabilities on top of the existing evidence-based Horse Intelligence Engine, without nuking your current scoring. I’ll design feature engineering pipelines (race cards, trackwork, past performances, horse history), train win/place and probability estimation models (XGBoost/LightGBM/Scikit-learn), then calibrate them (e.g., Platt/Isotonic) so confidence drops gracefully with missing data. Integration: ML outputs will feed into the current engine as additional, explainable evidence, every prediction remains traceable via feature/SHAP-style explanations and model cards. Deliverables will include production-ready ML pipeline, unit tests, model evaluation & calibration reports, and deployment docs for continuous retraining. To make this “reproducible evidence,” I’ll align training/validation with your current import pipeline and add data/versioning discipline in Git. Before we train anything, what exact inputs from your current evidence engine should ML be allowed to influence (and which must remain untouched), so we don’t accidentally replace proof with vibes? Similar work: sports analytics + probability-calibrated ranking models for prediction systems; portfolio links available on request. Timeline: ~30 days for v1 pipeline + integration + evaluation, then iterate as dataset grows.

@zahidhasta

Hi, Your approach of combining evidence-based analytics with machine learning is exactly how I prefer to build production ML systems. Rather than replacing your existing intelligence engine, I'll integrate explainable ML models that strengthen confidence while preserving the architecture you've already invested in. My experience includes Python, FastAPI, PostgreSQL, Pandas, NumPy, Scikit-learn, XGBoost, PyTorch, feature engineering, probability calibration, and production AI pipelines. I'll build reproducible feature engineering workflows from historical race data, train and validate prediction models for win/place probabilities, calibrate outputs, and integrate them cleanly into your existing backend with unit tests and documentation. I place strong emphasis on explainability, reproducibility, and confidence-aware predictions, ensuring missing or incomplete data lowers model confidence rather than producing unreliable results. I'm looking for long-term collaborations and would be glad to help evolve the intelligence engine as your historical dataset grows and new patterns emerge. Best regards, Zahid Hassan

@devanshkhandekar

Hi,I am a seasoned Applied ML Engineer(6+ yoe)& I can extend your existing Horse Intelligence Engine with reproducible ML predictions while preserving its rule-based evidence,factor ledger,confidence logic,FastAPI APIs,& PostgreSQL architecture Proposed Approach: -Data Audit & Feature Engineering:Analyze historical race cards & participant histories to engineer key predictive metrics (recent form,track suitability,weight,& opponent strength) -Modeling & Evaluation:Develop separate win/place probability & ranking models (LightGBM/XGBoost) using strict time-aware validation to prevent data leakage. Evaluate performance via Log Loss,Brier score,& ROC-AUC -Ensembling & Explainability:Blend ML probabilities with existing evidence scores,dynamically penalize confidence for missing data,& expose SHAP-based feature explanations -MLOps & Deployment:Implement automated retraining pipelines,model versioning,& FastAPI integration,supported by unit tests & comprehensive deployment documentation Relevant experience: -Developed sports/race computer-vision systems for finish-line analysis using detection,tracking,lane assignment,crossing timestamps,& ordered result generation. -Built predictive-maintenance intelligence engines combining ML risk scores with engineering rules,confidence bands,explanation traces,& audit logs closely aligned with your existing Factor Ledger -Developed production FastAPI pipelines with PostgreSQL-style persistence,modular inference services,retraining workflows

@KrisuTechkul

Krishna here from Delhi. With our proficiency in FastAPI, Python, and ML, we are well-placed to support the evolution of your Horse Racing Intelligence Platform. Our aptitude for Machine Learning, particularly Scikit-learn and XGBoost align perfectly with your project complemented by our expertise in Pandas, NumPy and PostgreSQL. The understanding and experience we have accrued in feature engineering, probability calibration, model validation and ranking systems specifically in sports and betting analytics will be invaluable for constructing a dependable prediction engine that stays grounded in evidence-based scoring. The projects we've successfully executed substantiate our deep-rooted command in the tools and technologies you require. For instance, our ability to develop an accurate demand forecasting model using ML algorithms, handle conversational AI proficiently or automating quality control using computer vision systems cogently showcases the range of our skills which we can efficiently deploy on your platform. A carefully crafted modelling approach blending time-series modelling with ranking and recommendation will be our compass for delivering custom-built horse racing intelligence tool that is rightfully yours.

@Web1devloper

As both a seasoned Data Analyst and Machine Learning Engineer, I have hands-on experience in delivering predictive analytics, which aligns perfectly with your project requirements. Over the course of my 8-year career, I've completed numerous ML projects using Python, Scikit-learn, and LightGBM, much like the ones you seek talents for now. I can design and implement sound machine learning models from horse racing historical data that come together seamlessly with your existing backend architecture. Furthermore, my knowledge of Pandas and NumPy is just what you need for efficient feature engineering pipelines construction and data handling in PostgreSQL. Your preference for maintaining model transparency and explainability resonates deeply with me - extracting meaningful insights from complex datasets and serving them in a digestible manner as a mainstay of my work. In terms of important notes, I strictly discourage hard-coded predictions, knowing that authentic ML outputs are crucial for your predictive racing platform.

@lalitdeveloper7

Hi, I can develop and integrate machine learning models into your existing Horse Racing Intelligence Platform while preserving the current evidence-based architecture. I have experience with Python, Scikit-learn, XGBoost, LightGBM, Pandas, NumPy, PostgreSQL, FastAPI, Feature Engineering, Time-Series Analysis, Probability Calibration, and Predictive Analytics. The solution will include feature engineering, prediction models for win and place probabilities, model validation, confidence calibration, explainable predictions, continuous retraining, and seamless integration with your existing backend. The code will be clean, reproducible, well-documented, and designed for long-term scalability. Please let me know further. Thanks.

@jaroslavcaprata

Hi! Hope you're doing well. Your Horse Racing Intelligence Platform is an excellent match for my ML and backend experience. Skilled in Python, FastAPI, PostgreSQL, Pandas, NumPy, XGBoost, LightGBM, feature engineering, probability calibration, and explainable ML models. I build reproducible, production-ready pipelines that integrate seamlessly with existing architectures. Looking forward to working with you. Thank you, Jaroslav Caprata.

@Vsion2

Hi there, I'm Alema, I can do the task perfectly and professionally please message me for a detailed discussion. I have completed 300+ projects with a 100% Positive Rating, If you are looking for Quality work, look no further. We are always available 24/7 to help employers without limitations, and delivery is guaranteed on time. Thank you!

@Hrolgar

Your brief is clear on one thing: the ML layer goes on top, not instead of, the existing confidence engine. That constraint shapes the whole design. I'd build the ML component as a standalone module with its own feature pipeline, XGBoost for win probability and LightGBM for place, calibrated with isotonic regression on held-out historical races. The outputs would be calibrated probabilities the Horse Intelligence Engine can weight alongside its current evidence scores, so it combines both signals rather than having to pick one. Feature engineering from race cards and trackwork pulls from what you're already storing in Postgres. One thing I'd want to nail down before firming up scope: how the existing engine expresses its output. A ranked confidence score vs a direct probability changes how I wire the integration layer. 37,500 INR / 21 days, single milestone. This is an indicative estimate from the brief; I'll give you a firm quote once I've seen the current schema and scoring logic.

@zain556

Hi, Your project is exactly the kind of ML challenge I enjoy—enhancing an existing intelligence engine with explainable, evidence-based machine learning rather than replacing it. I have experience building production ML solutions using **Python, FastAPI, PostgreSQL, Pandas, Scikit-learn, XGBoost/LightGBM, and Git**. I've developed feature engineering pipelines, predictive models, probability calibration, automated retraining workflows, and integrated ML into live backend systems. For your platform, I would: * Build robust features from race history, trackwork, race cards, and results. * Train and compare models for win/place probability prediction. * Apply probability calibration for reliable confidence scores. * Integrate predictions into your existing Horse Intelligence Engine while preserving your rule-based architecture. * Ensure every prediction is explainable, reproducible, and confidence decreases when data is incomplete. **Estimated Timeline:** 1–3 weeks for a production-ready pipeline, testing, documentation, and integration. I'm looking for a long-term collaboration and would be excited to help evolve the intelligence engine as your historical dataset grows. I'd be happy to discuss the data structure and recommend the best modelling approach for your platform. Looking forward to working with you!

@sgwtechnology

Machine Learning Engineer needed for an existing Horse Racing Intelligence Platform. Backend (FastAPI + PostgreSQL) is already built with rule-based analytics, and the goal is to add ML-driven prediction models without replacing the current evidence-based system. Scope: • Build ML models using historical race data • Feature engineering from race cards, trackwork, results • Win/place probability prediction models • Model evaluation, calibration, retraining pipeline • Integrate outputs into existing engine with explainability • Maintain reproducible, testable ML workflow Requirements: • Python (Pandas, NumPy, Scikit-learn / XGBoost / LightGBM) • Experience with ML pipelines and validation • PostgreSQL + FastAPI integration • Strong focus on explainable AI and data integrity Deliverables: • Production-ready ML pipeline • Model evaluation + documentation • Clean integration with existing backend • Deployment notes + tests Please include ML project experience, approach, timeline, and relevant portfolio.

@AcuteTech06

Hello, I have reviewed your Horse Racing Engineer project. I can help build a scalable horse racing data platform with PostgreSQL database design, data pipelines, API development, and performance-focused backend solutions. My expertise includes: ✅ PostgreSQL Database Design & Optimization ✅ Data Processing & ETL Pipelines ✅ API Development (FastAPI / Node.js) ✅ Data Ingestion from JSON/CSV Sources ✅ Backend Architecture ✅ Analytics Dashboard Support ✅ Automation & Custom Software Solutions I can help structure racing data such as horses, races, results, performance metrics, and prepare clean data systems for analytics and prediction models. I would like to discuss your requirements and deliver a reliable, scalable solution. Regards, Acute Tech Solutions

@PyMLExpert1

I am highly interested in joining your project as a Machine Learning Engineer to introduce data-driven learning capabilities into your ongoing Horse Racing Intelligence Platform. As a Python and Machine Learning expert, I have extensive experience building robust, explainable predictive models and real-time processing systems, making me an excellent fit to enhance your platform while strictly preserving your existing evidence-based architecture. ​Suggested Modelling Approach To introduce data-driven scoring without replacing your established, rule-based analytics engine, I suggest implementing a parallel ensemble or stacking architecture. We can develop probabilistic classification models—such as calibrated LightGBM or XGBoost models—specifically optimized for predicting win and place probabilities using historical race data. For feature engineering, I will design structured pipelines utilizing your existing production import pipelines, extracting robust temporal and categorical factors directly from race cards, trackwork, results, and deep horse histories. Crucially, to preserve explainability for every prediction, I will utilize SHAP or LIME frameworks to transparently break down model factors alongside your Factor Ledger. To address the data integrity requirement, the pipeline will be explicitly programmed so that missing data naturally dampens the model’s prediction confidence instead of fabricating values.

@iterativeTech

As an AI and Cloud Data Engineering Specialist, I bring a nuanced understanding of machine learning and the potential to drive measurable business outcomes. Having worked extensively in finance, healthcare, insurance, and enterprise environments, I am keenly aware of the importance not just of building scalable and production-ready systems but also of aligning them with your specific business objectives - which is exactly what I've done for diverse organizations. Having worked on past projects like building predictive models for classification and forecasting, recommendation systems, NLP, and generative AI applications, I'm well-versed in technologies like Scikit-learn, XGBoost, LightGBM, Pandas, NumPy - precisely what your project requires. My expertise also extends to PostgreSQL and FastAPI. In terms of preferred experience, I might not have direct involvement with sports analytics or horse racing data but my capabilities in time-series modeling and ranking recommendation systems can be readily leveragedfor this opportunity. To manage such an extensive process as described for your project- spanning from data engineering to feature engineering to model validation- granularity is paramount. This is where my skills intersect to create highly secure yet cost-effective ETL/ELT pipelines on platforms like AWS and Azure.

@AlexNdungu

Hi, I’m confident enough in solving this that I’m happy to dive in and start troubleshooting before you award the bid. Your existing FastAPI/PostgreSQL architecture is exactly the kind of environment I enjoy working with. I have experience building production ML pipelines in Python using scikit-learn, XGBoost/LightGBM, Pandas, NumPy, and integrating predictions into existing APIs while keeping models explainable and reproducible. For this project, I'd build a robust feature engineering pipeline from historical race data, develop calibrated probability models for win/place prediction, implement continuous retraining and evaluation, and integrate the outputs into your existing Horse Intelligence Engine without disrupting the current evidence-based scoring system. Every prediction will include confidence and supporting factors. I'm available to start immediately and would be happy to discuss the most suitable modeling approach for your historical dataset. I’m looking for a long-term collaboration and can deliver clean, documented, production-ready code with tests and deployment guidance. Looking forward to discussing the project.