Quantitative research · Market microstructure · High-performance systems
I build systems that turn market questions into measurable answers.
I came to quantitative finance from a tier-3 college, without the usual institutional access or network. A final-year internship at a quantitative trading firm changed how I read markets — less chart, more system: information, incentives, probability, execution. I have been building in that direction since.
I am the founder of IndiQuant, a quantitative research venture exploring how collective intelligence can produce a stronger understanding of Indian markets. Currently in private beta.
The public description stays high-level by design. The models, architecture, and strategy remain private.
Markets are not one problem. They are a chain — information, research, signal, portfolio, execution, risk — and the chain feeds back on itself. Two questions pull me back repeatedly.
How does distributed human intelligence become measurable? Good ideas are scattered across people and disciplines. I am interested in systems that can recognise useful insight without flattening what made it useful.
How much edge survives execution? A model can be correct and still lose money. Latency, liquidity, market impact, inventory, and risk decide whether theoretical edge ever reaches the P&L.
| Project | What it is |
|---|---|
| orderbook-reconstruction | Limit order book reconstruction in C++ |
| Derivative-Pricing | Black–Scholes pricing and a learned approximation of the volatility surface |
| polars-pandas-benchmarking | Reproducible Polars vs pandas benchmarks with memory profiling |
| FinNews-Sentiment-Analysis | Financial news turned into a measurable sentiment signal |
| Fundamental-Financial-Data-Scrapper | Company filings into research-ready datasets |
| JP-Morgan-Quant-Projects | JPMorgan quantitative research job simulation |
Python and PyTorch for research. C++ where latency matters. SQL and Postgres for structured evidence. Linux, Docker, and AWS when something has to run beyond a notebook.