Software Engineer building production-ready LLM products, multi-agent systems, document intelligence workflows, and AI developer platforms.

I have 2.5+ years of experience shipping end-to-end software across AI, automation, full-stack product engineering, and cloud infrastructure. My recent work includes multi-provider LLM routing, context engineering, MCP integrations, restorable agent checkpoints, workflow automation, and Kubernetes-based deployment systems.

These days, I focus on making AI systems reliable, cost-efficient, and useful in production - from the agent and API layers to the infrastructure that keeps them running.

Feb 2026 - Present · Noida, India

• Designed a 4-agent orchestration system spanning planning, code generation, validation, and deployment.
• Reduced AI inference costs by 50% through context engineering, prompt restructuring, context pruning, and token-budget optimization.
• Built a 3-provider LLM orchestrator across OpenAI, Anthropic, and Gemini with automatic failover.
• Created restorable agent checkpoints with code-state tracking, diff inspection, rollback APIs, and editor integrations.
• Developed a multi-user MCP integration layer with OAuth 2.0 + PKCE, encrypted token storage, and streaming tool execution.
• Implemented custom-domain infrastructure with DNS verification, automated TLS, Kubernetes ingress, and cert-manager.

Jun 2024 - Jan 2026 · Noida, India

• Built a 5-stage document automation pipeline connecting email ingestion, OCR, LLM extraction, Excel export, and QuickBooks.
• Eliminated manual invoice data entry and built direct synchronization with accounting transactions.
• Developed scheduled email workflows using Celery Beat with hourly, daily, and custom cron execution.
• Led the migration from Docker Compose to Kubernetes using Helm, HPA, and production health checks.
• Built Redis-backed notification infrastructure with asynchronous workers, reusable templates, and automated tests.

I build systems where specialized agents can plan, generate, validate, and deploy applications through context-aware handoffs. My work includes multi-provider orchestration across OpenAI, Anthropic, and Gemini, automatic failover, context-window optimization, tool integrations, and restorable checkpoints.

I have built automated pipelines spanning email ingestion, OCR, LLM-based structured extraction, spreadsheet export, and QuickBooks synchronization - turning raw business documents into validated accounting workflows with no manual data entry.

A multi-tenant AI voice SaaS for generating speech from system and custom voices. Built with FastAPI, Next.js, tRPC, PostgreSQL, Prisma, Chatterbox TTS, and cloud object storage.

• Text-to-speech generation with configurable voice parameters
• Custom voice uploads and organization-based access control
• Google authentication, scoped voice libraries, and generation history
• Typed API integration from generation through audio storage

A MERN learning platform supporting course discovery, enrollment, video-based learning, progress tracking, instructor publishing workflows, payments, and ratings.

• Student and instructor experiences with role-based authorization
• Course, section, subsection, video, pricing, and publishing workflows
• JWT authentication, OTP verification, password recovery, and Razorpay payments

AI / LLM   Multi-agent systems · Context engineering · RAG · MCP · Tool calling · Structured outputs · OCR · FAISS · Chroma · Qdrant · SentenceTransformers

Backend   Python · FastAPI · Django · Node.js · Express.js · tRPC · SQL

Frontend   React · Next.js · TypeScript · Redux Toolkit

Data & messaging   PostgreSQL · MongoDB · Redis · RabbitMQ · Celery

Cloud & DevOps   AWS · DigitalOcean · Docker · Kubernetes · Helm · CI/CD · Autoscaling · Observability

• Reduced LLM inference costs by 50% through prompt restructuring, context pruning, and token-budget tuning.
• Built a 3-provider LLM orchestrator with automatic failover for stronger runtime resilience.
• Designed a 4-agent application-generation workflow spanning planning, code generation, validation, and deployment.
• Led a Docker Compose-to-Kubernetes migration with Helm, HPA, and production health checks.
• Built OAuth 2.0 + PKCE integrations with encrypted token storage and streaming MCP tool execution.

📍 Noida, India

_{Interested in agentic systems, production AI, and developer infrastructure? Explore my repositories or reach out - I'm always happy to connect.}