[MENTEE] Kushvinth #74
Description
Mentee:
Kushvinth Madhavan / Software engineer & AI systems builder working on ML, agents, and production AI infrastructure / gh: https://github.com/kusvhinth / Based in India, interested in probabilistic ML, time-series modeling, and scalable ML system design.
Mentor(s):
Why did you join sktime's mentorship program?
I joined the sktime mentorship program to learn how large-scale machine learning libraries are designed, maintained, and evolved in practice. I’m particularly interested in understanding how theoretical models for forecasting and probabilistic prediction are translated into consistent estimator APIs, reusable components, and maintainable architecture within the sktime ecosystem.
I also want to gain hands-on experience contributing to a mature open-source ML project and learn the workflow, testing discipline, and design decisions required for production-quality scientific software.
What topics are you working on?
- Understanding sktime’s estimator architecture, tags system, and pipeline composition
- Exploring probabilistic forecasting interfaces and prediction workflows
- Reviewing beginner-friendly issues and documentation gaps to start contributing
- Studying existing PRs/issues to understand contribution patterns and code standards
What are your learning goals?
- Build strong intuition for extensible ML API and estimator design
- Learn best practices for contributing to large open-source Python ML libraries
- Improve skills in testing, documentation, and code review processes
- Understand architectural trade-offs in real-world ML frameworks
- Make meaningful contributions that align with sktime’s long-term design principles
What's next for you after the mentorship program?
I plan to continue contributing to sktime beyond the mentorship period, gradually taking on more complex issues and architectural discussions.
Long term, I want to apply these learnings to build reliable ML/AI systems and contribute actively to open-source infrastructure projects in the machine learning ecosystem.