Hire the Best Python Numpy Developers
in South Africa

AI has made code easier to produce. It has not made quantitative research easier to trust. A backtest can be generated quickly. Building a research system that produces reproducible experiments, tests generalization, handles large parameter spaces, integrates machine learning correctly, and carries strategy logic safely toward live execution is a different problem. That is the work I specialize in. I design and build serious quantitative research, backtesting, optimization, and trading systems in Python, with particular depth in: VectorBT Pro and Numba Optuna and large-scale hyperparameter optimization Walk-forward testing and chronological cross-validation Robustness analysis and experiment reproducibility Machine learning for trading Market-data infrastructure Signal generation and live trading workflows Research platforms and operational tooling My work is built around the Quantforge Research Engine, my proprietary quantitative research framework. The research engine extends VectorBT Pro into a complete research environment covering the lifecycle around the simulation engine: Define → Validate → Queue → Source data → Simulate → Optimize → Stress-test → Analyze → Reproduce → Generate signals → Trade This gives suitable projects a substantial head start. Instead of billing you to rebuild job queues, data caching, experiment tracking, optimization orchestration, result analysis, ML integration, and trading infrastructure from scratch, I can focus development on what is actually specific to your research. That may include: Your strategy and portfolio logic Custom indicators, signals, stops, sizing, and execution rules Proprietary investment universes and market-selection policies Exchange, broker, or data-provider integrations Custom optimization objectives and search spaces Walk-forward and robustness methodology ML features, labels, models, and decision pipelines Research dashboards and diagnostic tooling Signal generation and production trading workflows The Quantforge Research Engine already supports configuration-driven experiments, multi-source market data, gap-aware caching, multi-timeframe datasets, persistent Optuna studies, smart and grid search, nested optimization, cross-validation, robustness scoring, MLflow model tracking, scalable result analysis, experiment reproduction, queue-based execution, live signals, and persisted trading state, and much more. My primary stack includes Python, VectorBT Pro, NumPy, Pandas, Polars, Numba, Optuna, CatBoost, XGBoost, TensorFlow, Scikit-learn, MLflow, FastAPI, Streamlit, PostgreSQL, MySQL, Redis, Docker, REST APIs, and WebSockets. I am a strong fit if you are seeking: A serious backtesting and quantitative research platform A systematic strategy requiring rigorous optimization and validation An ML-assisted trading research pipeline Infrastructure connecting research, signals, and execution A platform that must support repeatable experiments across strategies, markets, or clients A system that has outgrown notebooks and disconnected Python scripts I am unlikely to be the right developer if you need: A small standalone indicator A quickly assembled trading bot Guaranteed returns or a “profitable AI bot” A disposable prototype with no concern for reproducibility Someone to implement instructions without questioning the research design For a detailed look at the research infrastructure I bring to suitable projects, see the Quantforge Research Engine guide in my Upwork portfolio. I care about whether a strategy has been tested honestly, whether its results can be reproduced, whether optimization is discovering signal or exploiting noise, and whether the live system behaves like the backtest that justified it. If that is also how you think about quantitative development, we should talk.

I am a developer with some experience tinkering with small projects. I have prior experience in various programming languages - namely: - C - at a low intermediate level (with familiarity with tools like Valgrind (memory management), GDB (debugging), clang-format (code formatting)) - Python - at a advanced level, with familiarity with package management tools like poetry, environment management tools like pip-env, and code formating tools black/flake8/ruff, as well as having a beginner-intermediate level understanding of multiple mathematics/data science/machine learning frameworks namely - Numpy, Pytorch and Pandas, and experience web-scrapping with BeautifulSoup and Selenium. - Java - at an intermediate level - Bash - basic beginner level, with some basic commandline scripting experience Additionally, I have experience working with Microsoft's Office Suite - namely PowerPoint, Word, and Excel, so I am capable of preparing and presenting presentations if needed.

I have a BSc (Honours) degree, with majors in Computer Science and Applied Mathematics. I have had Upwork Top Rated status for more than 5 years, Top Rated Plus status (great success with bigger clients and longer contracts) for more than 2 years. I have an Upwork Job Success Score that is still 100% after 57 jobs, 10,000+ hours and 10 years full-time. I am a Certified Microsoft Professional in the field of Data Science. I have real technical skills, such as Python, ML & Deep Learning, Power BI, Excel, Tableau, VBA, SQL, Mongo and PHP, and I am certainly not afraid of getting my hands dirty, but I also like (and am able) to operate on a higher level, to see a project or requirement through from concept to completion in the most elegant and efficient manner. 1. Data Science skills include: Python, Numpy, Pandas, scikit-learn, Matplotlib, Plotly, Power BI, Tableau, Excel and ElasticSearch. 2. My technical and programming skill-set includes Python, Django, Flask, VBA, Excel, Access, PHP, SQL, mySQL, Jquery, Javascript, JSON, Ajax, HTML and CSS. 3. Cloud infrastructure skills: AWS, Azure, GCP. 4. Database skills: SqlServer, MySQL, PostGresSQL, MongoDB, BigQuery. 5. PM and BI skills include Power BI (Microsoft Certified), DAX, Power Query, Tableau, MS Project, PowerApps, MS Flow, SharePoint and Trello. 6. Peripheral skills include, amongst others, Knime, Git, & Vim.. I have great people skills, a proven track record in effective project management, the ability to work in an unstructured environment and be self motivated, and good relationship building and writing skills. I have more than 20 years of experience in real world web & application development and project management. I believe in ethical business practice, and my commitment is to provide online applications and IT business solutions of the highest possible quality.

I am an Senior Software Developer and part time university student currently pursuing a Bachelor of Science in Computer Science and Mathematics seeking to contribute to your Software Development department through a full time position. I have worked at LottoStar (1609 Group) beginning as a Junior Developer in February 2019 and advancing to Intermediate in June 2020. I attribute my quick rise to willingness to learn and work above and beyond my given requirements. I started out primarily on the backend working with Python, Django and Postgres and then later PHP and Phalcon. Once we expanded Expanded our codebase I shifted towards both backend and frontend development using HTML, CSS and pure Javascript at first and then later the React framework. This improved my skills using and designing APIs on the back and front end. One of our most difficult tasks, which I was a primary member of, was upgrading our software on the backend as our versions of Python and Django were deprecated.