projects
Applied ML systems across visual search, responsible AI, personalization, and multimodal reasoning.
Selected systems
Projects that connect model quality with real deployment constraints.
My strongest work sits at the intersection of retrieval, ranking, multimodal modeling, responsible AI evaluation, and production ML infrastructure. These projects emphasize measurable impact: relevance, latency, faithfulness, robustness, and user-facing outcomes.
Visual & Multimodal Retrieval System
Vision-text retrieval with FAISS, learned reranking, Ray Serve latency optimization, and responsible AI evaluation
- 1M+ vectors indexed
- 8-12% Recall@100 gain
- 15-20% lower p95 latency
User Behavior Segmentation & Predictive Profiling
Behavior modeling pipeline for recommender systems using clustering, anomaly detection, sequence classification, and drift monitoring
- Clustering + anomaly detection
- Transformer/LSTM/XGBoost benchmarks
- Drift monitoring harness
ICICI Bank Personalization & Ranking Platform
Production ML recommendation and ranking systems serving 1M+ daily requests with measurable CTR and relevance gains
- 1M+ daily requests
- 9% CTR lift
- 12% relevance improvement
Verified Medical NLP – RDF-Grounded Jamba RAG
RDF-grounded medical question answering with deterministic hallucination checks and Jamba MoE reasoning
- 14% lower hallucination rate
- Claim-level verification
- Auditable RDF grounding
EDI-Summ: Multimodal Clinical Summarization
State-of-the-art multimodal summarization architecture for clinical documents
- ACL 2024 long paper
- 0.81 factual recall
- Vision cross-attention
Gestural AI – Real-Time ASL Interpreter
A real-time American Sign Language recognition system achieving 94% accuracy
- 94% recognition accuracy
- 20K+ videos
- Real-time inference
Earlier research and engineering
I also keep older work on clinical summarization, accessibility, climate decision support, and open-source software because it shaped how I think about rigorous evaluation and practical systems.