I turn rigorous math into production AI — content discovery at OneScrybe, sports analytics at SportsAnalytik, and a lifetime of problems worth solving.
I'm a data engineer and applied-ML practitioner with a thirty-year track record of shipping systems that hold up under real load — from terabyte-scale Elasticsearch architectures to retrieval-augmented LLM pipelines.
My foundation is mathematics. Dual master's degrees in Math and Computer Science let me move between deep theory and hands-on engineering: architecting AI platforms, building UI frameworks and custom visualizations, and translating dense computational ideas into clear business value.
Today I focus that on two products of my own, alongside consulting in data science and search. When I'm not building, I'm usually somewhere in sport, music, technology, or food — the four things I keep coming back to.
The numbers behind the game. I build dashboards and models across basketball, football, cricket, and tennis — where statistics meet a great rivalry.
A deep love of Indian classical music alongside an ever-running playlist. Sitar, tabla, and shehnai sit comfortably next to whatever's on rotation this week.
NLP, retrieval-augmented generation, vector search, and the engineering to make them real. Python, Angular, AWS — and a standing curiosity about what's next.
Cuisine as a system to explore — flavor, region, technique. The same taxonomic curiosity I bring to data shows up at the dinner table.
An AI-powered content discovery platform and creator toolkit built on top of YouTube — spanning music, education, travel, food, and studio verticals, with multilingual semantic search at its core.
A sports analytics platform turning raw play data into clear, interactive insight — player comparisons, shot charts, and gap analysis across the NBA, NFL, WNBA, cricket, and tennis.
Leading technical strategy across two AI ventures — pairing strong mathematical foundations with production-level AI engineering to build high-performance, scalable platforms.
Assessed and implemented large language models (LLMs) to optimize daily workflows within classified projects. Working on Retrieval Augmented Generation (RAG) using python, fastapi, typescript that takes documents from elastic search in chunks, analyzed by LLM and consolidating the results to answer user question.
Additionally, mapped elasticsearch fields across the assets to establish dependency relationships, enabling precise impact assessments for planned schema or field modifications.
Built and deployed end-to-end data systems on Cloudera Hadoop and Spark. Orchestrated cross-device match testing that generated $5M/year in growth and automated ETL for major customers.
Led product development and architecture, synthesizing clinical and FDA data into a MapReduce system. Applied regression, random forests, text mining, and NLP across the data lifecycle.
Managed engineering for an ISP provisioning platform. Built the proof of concept that secured a first round of $10M in venture funding; company acquired in 2003.
Let's build something worth solving.
Open to consulting in data engineering, applied ML, and search — and always up for a good problem.
Multilingual music discovery and analysis — semantic search across Indian classical and global catalogs.
kirthiraman.com/music →NFL record-book search and sports analytics — a generative AI demo over a structured records dataset.
kirthiraman.com/nfl →RAG systems, local LLM pipelines, and applied ML experiments — the engineering notebook behind the work.
kirthiraman.com/tech →Problems, proofs, and visualization — the mathematical foundation that drives everything else here.
kirthiraman.com/math →