AI Engineer at VSky Solutions, working across agentic AI, retrieval-augmented generation, and full-stack backend architecture. I ship systems that route, verify, and recover — not demos that fall over outside a notebook.
I'm a computer science student (graduating 2026, D.Y. Patil Agriculture & Technical University) working as an AI Engineer Intern at VSky Solutions, where I design and ship production systems across generative AI, RAG, and agentic workflows.
My work sits at the intersection of backend architecture and LLM orchestration — building the routing logic, verification layers, and failure handling that separate a working prototype from a system people actually rely on. I've had architecture proposals adopted into company products, restructured legacy tools into layered, maintainable codebases, and built a handful of personal systems that push further than the day job allows.
Outside of assigned work, I build things that meet the same bar I hold enterprise tools to — multi-agent research systems, evaluation pipelines, and full applications from database to UI.
A mix of shipped company systems and independent projects — chosen for what they actually demonstrate, not for volume.
A retrieval-augmented Q&A system that answers questions directly from internal SOP documents — shipped at VSky, now getting a more advanced personal successor.
A production HR screening platform delivered as a Microsoft Outlook add-in — migrated from fragile n8n webhook chains to a custom Google ADK multi-agent system.
An enterprise-grade document intelligence platform — hybrid retrieval, cross-encoder reranking, answer verification, and a closed feedback loop that improves future retrieval and prompting quality.
A full-stack clinical intelligence system with a 6-agent LangGraph pipeline, integrating real healthcare APIs for trial matching, drug-interaction checks, and physician report generation.
A 6-stage deterministic multi-agent pipeline that generates, scores, selects, and evaluates professional business strategy reports.
A full-stack lineage tool that connects SAP BusinessObjects universe objects to their underlying physical database structures — in use by clients, not just internally.
Multi-agent ERP automation — Outlook email to AI extraction to validation to ERP insertion, end to end.
YOLO-based object detection pipeline that locates receipt fields and auto-inserts extracted data into ERP systems.
Self-architected multi-RAG hosting platform with login and dashboard, built after proposing the design at work.
A persistent multi-agent system simulating a full software-engineering org, producing real code against a real repo.
Fine-tuned Qwen2.5-3B (LoRA) served via Ollama, with an SSE-streaming backend and intent-based routing across six task types.
A virtual classroom platform with multi-role auth, S3 file storage, and 35+ passing Pytest suites.
A VGG-16-based nail disease screening app classifying 17 conditions, with a custom training pipeline.
Not a tag cloud — the layers a request actually passes through, top to bottom.
Looking for AI Engineer, Agentic AI Engineer, or Full-Stack AI roles where production reliability actually matters.