Associate Full Stack AI Engineer with 3+ years shipping enterprise AI systems in production. I design and deploy LLM-powered applications — from AI agents and RAG pipelines to the cloud-native infrastructure that runs them at scale.
I turn ambiguous business requirements into reliable, measurable, production-grade AI systems.
| Project | Stage |
|---|---|
| WebFoxShield — AI Cybersecurity Platform | Production |
| Pickup Express — Enterprise Logistics Platform | Production |
| CONDUCTOR — Hybrid AI Gateway for Dynamic LLM Routing | AMD Hackathon |
| CareBank — AI Digital Wealth & Lending Platform | IDBI Hackathon |
| MCP Logistics Assistant — Slack AI Agent | Slack Challenge |
| AI Voice Generation Platform — Neural TTS Research | Research |
| Domain | Focus Areas |
|---|---|
| AI Systems | Enterprise AI, LLM Engineering, AI Agents, LangGraph, RAG, Prompt Engineering, Vector Databases |
| Architecture | Distributed Systems, Microservices, System Design, Event-Driven, Cloud-Native |
| Backend | REST APIs, FastAPI, Flask, Django, Node.js, Express.js, PostgreSQL, Redis, MongoDB |
| Infrastructure | Docker, Kubernetes, Terraform, Ansible, GitHub Actions, Jenkins, ArgoCD, Nginx |
| Observability | Prometheus, Grafana, Structured Logging |
| MLOps / LLMOps | Model Serving, Prompt Management, CI/CD for LLMs, Evaluation Pipelines |
AI-powered Cybersecurity Platform — Real-time web threat detection at enterprise scale.
Business Problem: High false-positive rates in web threat detection; inability to scale analysis across enterprise traffic volumes
Architecture: Event-driven microservices with async ML inference pipeline; multi-model ensemble for threat classification
Key Features: Real-time request inspection, ML-based anomaly detection, automated incident response, dashboard analytics
Tech Stack: Python, FastAPI, PyTorch, TensorFlow, Redis, PostgreSQL, Docker, Kubernetes, AWS
Enterprise Logistics Platform — Last-mile pickup orchestration and route optimization.
Business Problem: High dispatch latency and suboptimal courier allocation in last-mile logistics networks
Architecture: Distributed services with geospatial routing engine; event-driven dispatch queue
Key Features: Real-time courier tracking, geospatial route optimization, automated dispatch, SLA monitoring
Tech Stack: TypeScript, Node.js, Express.js, PostgreSQL, PostGIS, Redis, RabbitMQ, Docker, AWS
Hybrid AI Gateway for Dynamic LLM Routing — Multi-agent workflow orchestration.
Business Problem: Coordinate multiple specialized AI agents reliably at scale while optimizing cost and latency across LLM providers
Architecture: LangGraph-based agent graph with stateful task queue; hybrid routing layer for multi-provider LLM dispatch
Key Features: Multi-agent orchestration, dynamic LLM provider routing, cost optimization, fallback chains, observability
Tech Stack: Python, LangChain, LangGraph, FastAPI, Redis, PostgreSQL, Docker, Kubernetes
AI Digital Wealth & Lending Platform — Secure, compliant AI-powered financial services.
Business Problem: Surface accurate, compliant answers from sensitive financial domain data while meeting regulatory requirements
Architecture: RAG pipeline with secure multi-tenant isolation; vector search for document retrieval at scale
Key Features: Secure document retrieval, compliance-aware RAG, multi-tenant data isolation, audit logging
Tech Stack: Python, FastAPI, LangChain, pgvector, MongoDB, Redis, Docker, AWS
Neural TTS & Voice Synthesis — Low-latency voice generation at scale.
Business Problem: Deliver real-time, natural-sounding voice synthesis with sub-500ms latency across concurrent streams
Architecture: GPU-accelerated inference workers behind streaming audio gateway; load-balanced worker pool
Key Features: Neural TTS, voice cloning, streaming audio, concurrent request handling, model hot-swapping
Tech Stack: Python, PyTorch, WebSockets, Docker, Kubernetes, AWS
Generative AI Sales Platform — Automated webinar content and sales follow-up.
Business Problem: Manual end-to-end webinar creation and sales follow-up is slow, inconsistent, and does not scale
Architecture: LLM orchestration with event-driven content pipeline; multi-stage generation with human-in-the-loop
Key Features: Automated script generation, slide creation, personalized follow-up emails, analytics dashboard
Tech Stack: TypeScript, Next.js, Python, FastAPI, OpenAI, Redis, Docker, AWS
MCP-based AI Assistant — Conversational logistics assistant integrated with Slack.
Business Problem: Logistics teams need instant access to shipment data, ETAs, and operational insights without switching contexts
Architecture: MCP server exposing logistics APIs; Slack Bolt integration for conversational interface
Key Features: Natural language shipment queries, real-time ETA lookups, proactive alerting, multi-modal responses
Tech Stack: Python, MCP SDK, Slack Bolt, LangChain, Redis, PostgreSQL, Docker
Event-Driven Architecture, Distributed Systems Design, Workflow Orchestration, Real-Time Data Processing, API Design, Containerization, CQRS, Data Pipeline Architecture
Programming
Frontend
Backend & API
AI & Machine Learning
Databases & Message Queues
Cloud & Infrastructure
CI/CD & Automation
Observability
Tools
- Advanced LLMOps — Model evaluation, prompt management, and CI/CD for LLMs at production scale
- Distributed AI Systems — Horizontally scalable multi-agent architectures
- Multi-Agent Systems — Complex agent coordination patterns and tool-use optimization
- Kubernetes (Production) — Advanced orchestration, auto-scaling, and service mesh
- Large-Scale AI Infrastructure — GPU cluster management and inference optimization
