AI Engineer · Distributed Systems · Full Stack · B.E. AIML @ DSATM Bengaluru
I design and ship production-grade systems — LLM-powered RAG pipelines, fault-tolerant distributed caches, and real-time computer vision stacks — with an emphasis on correctness, scalability, and measurable engineering impact.
Chunk-aware RAG delta-sync engine cutting embedding API calls by >90% with sub-50ms search latency across sharded gRPC worker nodes and custom KD-Trees.
Fault-tolerant in-memory cache with O(1) latency via custom thread-safe HashMap, master-slave replication, and monotonic epoch split-brain prevention.
Secure code execution with Docker-isolated sandboxing, async Celery job pipeline, and local LLM (Qwen 2.5 Coder) for automated static analysis and complexity reviews.
Clinical eye segmentation with sub-100ms inference via quantized ONNX export. Mean detection error within 2mm on live video streams with temporal ROI smoothing.
Real-time harassment detection using YOLOv9 + SORT tracking + ONNX-optimized LSTM classifier over 30-frame behavioral windows targeting 85%+ confidence.
High-fidelity medical RAG using LangChain, ChromaDB, and Groq (Llama 3.3) grounded in 2025–2026 clinical guidelines for accurate, citation-backed responses.
Medical Report Simplifier (LLM + Django); outperformed all 70 competing teams.
"A Review on AI-Enabled Wildlife Preservation and Management System" — IEEE International Conference on Next-Gen Quantum & Advanced Computing.
Institutional Nominee — selected from 100+ teams for innovative AI/ML technical solution.
Delivered "AI as Boon or Bane" exploring societal, ethical, and technical impacts of AI before a university audience.
Open to AI/ML engineering roles, backend & distributed systems, research collaborations, and open-source contributions.