I design and deploy production-grade AI systems—from multi-agent orchestrators and multimodal RAG to edge computer vision. I focus on the entire lifecycle: data pipelines, latency optimization, and deterministic reliability.
4
Production Systems
E2E
Production Delivery
Low Latency
Edge Optimized
Deterministic
Reliable AI
Currently Building
- Retail Intelligence Platform
- Multi-Agent ERP Assistant
- Multimodal Knowledge Engine
AI / ML
- Python
- OpenVINO
- FAISS
LLMs / RAG
- Qdrant
- Custom Agents
Backend
- FastAPI
- Redis
- AWS S3
Deployment
- AWS EC2
Engineering philosophy
I believe AI is a tool to solve business problems, not just a science experiment. My approach starts with the operational bottleneck—whether that's manual RFP evaluation, retail footfall tracking, or ERP data retrieval—and builds an AI architecture specifically engineered to solve it.
Shipping an AI demo is easy; shipping a production system is hard. I focus on the unglamorous engineering that makes AI reliable: deterministic regex overlays to prevent LLM hallucinations, FAISS margin gates to prevent false biometric matches, and context densification to optimize token budgets.
Problem First
We start with real operational bottlenecks, not technology.
Reliability First
Deterministic guardrails ensure consistent and trustworthy AI.
Production Focused
Architectures designed for scale, latency, and cost-efficiency.
Business Impact
Every system is measured by the value it creates.
Selected work
Production AI systems built for real-world deployment across computer vision, multimodal AI, and enterprise automation.
Multimodal Knowledge Engine
Indexes docs, images, and web content — then answers with cited sources over REST and voice.
Impact
Sub-200ms hybrid retrieval + rerank
Procurement AI
AI vendor proposal evaluation that scores compliance, extracts commercial terms, and ranks winners.
Impact
Automates hours of manual RFP review
Retail Footfall Analytics
Multi-camera edge CV that turns RTSP feeds into footfall, attendance, demographics, and uniform compliance.
Impact
CPU load ~60% → ~15% per stream
Retail AI Orchestrator
Multi-agent retail chatbot that routes plain-English questions to domain agents pulling live .NET ERP data.
Impact
Routes queries across 10+ ERP domains
Experience
Where I've applied AI and engineering to build, automate, and solve real-world problems.
Associate AI Engineer · DigixValley
PresentSep 2026 – Present
- Designing AI system architecture for production-ready solutions across client and internal products.
- Building virtual try-on experiences that combine computer vision and generative models for real shopping workflows.
- Delivering AI solutions for real-world problems — from model selection and pipelines to deployment and iteration.
Tenure
Ongoing
Full-time
AI Engineer · TechReFi
EndedSep 2025 – Sep 2026
- Built a RAG platform indexing manuals and procurement records into Qdrant with hybrid search — now powers automated RFP/RFQ extraction on AWS EC2.
- Combined metadata filtering with cross-encoder reranking to cut search time to under 200ms.
- Shipped real-time CCTV analytics (OpenVINO INT8) cutting CPU load per stream from ~60% to ~15% while sustaining 30+ FPS across 8 cameras.
Impact: Real-time performance, cost optimized, production ready.Intern → Full-time in 3 months
Tenure
1 year
Full-time
AI/ML Engineer Intern · CodeCelix
InternAug 2025 – Nov 2025
- Built a RAG chatbot (Qdrant, GPT-4) that automated common support queries.
- Led a team of 4 interns on a CV parsing pipeline (spaCy + BERT) that improved recruiter shortlisting.
Tenure
3 months
Internship