Hi, I'm
Rishi Raj Grandhe
ML engineer & founder building AI systems that serve millions — from national disease surveillance to TB care at scale.
Exploring what's next — open to founding-team and senior ML roles.
About
Building AI that works at scale
I'm an ML engineer and repeat founder who's spent the past five years deploying AI systems that reach millions — from a national disease surveillance pipeline scanning 400M+ media sources for India's NCDC, to a TB treatment adherence model covering 100K+ patients across 15 states. I've published at ACL and NeurIPS, built products from zero to pilots, and I'm always looking for the next hard problem where ML can create real-world impact.
Full experience →Research
Published work
Peer-reviewed papers at top ML and NLP venues.
Health Sentinel: An AI Pipeline For Real-time Disease Outbreak Detection
A national-scale, real-time, multi-stage and multi-lingual disease outbreak detection pipeline deployed with India's NCDC — scanning 400M+ media sources, issuing 100K+ alerts, and verifying 5K+ outbreaks through the Media Scanning and Verification Cell.
Read case study →Predicting Treatment Adherence of Tuberculosis Patients at Scale
A similarity-encoded ensemble model deployed across 15 Indian states that flags high-risk TB patients likely to drop off treatment — covering 100K+ patients with fairness-aware reweighting to ensure equitable recall in underserved districts.
Read case study →Projects
Things I've built
Products, research prototypes, and experiments.
Pulp AI
A simulation platform that lets product teams test copy, features, and pricing against AI-generated customer personas — shipping decisions in minutes instead of weeks. Built on FastAPI, Postgres (pgvector), Redis, and Docker.
Pest Detection System
Trained EfficientNet, ResNet, and Inception models for pest identification on cotton and rice farms. Advanced data augmentation improved pipeline robustness by 14%, directly helping 10,000+ farmers take corrective action without compromising yield.
CPI Forecasting across Indian Markets
Scraped ~1.2M CPI price points from 300+ Indian markets and trained LSTM models that reduced 30-day-ahead MAE by 18% versus ARIMA, revealing inflation hot-spots across eight economic zones.