$ ./connect --all
[madhavendranath@arch ~]$ cat about.md
Third-year B.Tech student in AI & Data Science at IIIT Kottayam, working on machine learning, web development and systems engineering. Co-author on a published research paper in Biomedical Physics & Engineering Express (2025) on temporal patient trajectory modelling using LSTM autoencoders on EHR data.
I like building things that are technically interesting end-to-end — from RAG pipelines with hybrid retrieval and cross-encoder re-ranking, to LSTM-based portfolio optimisers with custom Sharpe-ratio losses, to compiler/linker projects in C and Rust.
[madhavendranath@arch ~]$ cat current.log
- Building RAG systems with hybrid retrieval (pgvector + BM25), RRF, and cross-encoder re-ranking
- Deep-learning-based portfolio optimisation (PyTorch, LSTM, FinBERT)
- Systems projects — an ELF static linker and an LLVM-based mini C compiler targeting NVPTX
- Healthcare ML research on longitudinal EHR data (MIMIC-IV)
- Production full-stack apps with Next.js, FastAPI, and PostgreSQL
[madhavendranath@arch ~]$ cat stack.md
// ai · ml · data
// languages
// backend · infra
// frontend
// databases
// mobile
[madhavendranath@arch ~]$ cat responsibilities.md
- Machine Learning & Research: Published research in healthcare ML; building models on time-series, NLP, and computer vision tasks with PyTorch and scikit-learn.
- Retrieval-Augmented AI: Production RAG pipelines with dense + sparse hybrid retrieval, Reciprocal Rank Fusion, cross-encoder re-ranking, and citation-grounded generation.
- Quantitative Development: Deep-learning models for portfolio optimisation and trading research — custom losses (Differential Sharpe Ratio), walk-forward backtesting, CVaR constraints.
- Full-Stack Engineering: End-to-end apps with Next.js, FastAPI, and PostgreSQL — from schema design to deployment.
- Systems & Compilers: Exploring low-level work — ELF linkers, LLVM-based compilers, NVPTX code generation.
[madhavendranath@arch ~]$ top-langs --username madhav-000-s
[madhavendranath@arch ~]$ ./connect --help
Open to internships, research collaborations, and quant / ML / SWE opportunities.