💻 Machine Learning & Systems Enthusiast 🚀 Open Source Contributor (sktime ecosystem) 🧠 Focused on Time Series & Agentic Workflows
Languages & Core Python • SQL • Git
Machine Learning NumPy • pandas • scikit-learn • SHAP Time Series: sktime
Systems & Backend FastAPI (basic) • ML pipelines • Evaluation systems
Tools Git • GitHub • VS Code • Jupyter
| PR | Repository | Focus |
|---|---|---|
| [ENH] Agentic workflow benchmark suite #396 | sktime-mcp | Benchmarking framework |
| [FIX] Substring fallback for tag suggestions #366 | sktime-mcp | Utility + UX improvement |
| [BUG] Fix TimesFMForecaster Python version bounds #10106 | sktime | Dependency + compatibility |
| [BUG] Fix HierarchicalProphet dependency bounds #10103 | sktime | Version control fix |
| [DOC] Modernize SHAP API in notebook #4926 | shap | API modernization |
Building a benchmarking framework for sktime-mcp to evaluate agent-generated pipelines across:
- Pipeline validity
- Task alignment
- Predictive performance
📌 Proposal: https://github.com/Rohitstwt/esoc-2026-sktime-agentic-benchmark
- Time series systems (sktime internals)
- Agentic workflows & evaluation
- Open-source ML infrastructure
- Applied ML + explainability
I focus on understanding systems deeply — fixing real issues, improving reliability, and building structured solutions like evaluation frameworks rather than just using libraries.
📫 Open to collaborating on meaningful ML systems and open-source projects.