Backend & AI Systems Engineer · Pune, India
Building intelligent agent infrastructures, low-latency trading systems, and production-grade AI automation

I'm a Backend & AI Systems Engineer with ~2 years of experience building production-grade intelligent systems. Currently working at GoQuant (Remote, Miami, USA) on:

• 🧠 Modular MCP (Model Context Protocol) frameworks — enabling AI agents to securely debug and operate in high-frequency trading environments
• ⚡ Low-latency Smart Order Routing (SOR) — real-time multi-venue liquidity evaluation using Python + C++
• 📊 Consolidated Level-2 Order Book — real-time aggregation across multiple exchanges with <0.1ms overhead
• 🤖 "GoOracle" — an intelligent trading operations assistant with agentic AI workflows (LangChain + LangGraph)

Previously at Codeinbound LLP as a Full Stack Developer, building enterprise NMS (Network Management Systems) with real-time dashboards, SSH/RDP connectivity, and SNMP-based monitoring.

Remote | Miami, USA | High-Frequency Trading Tech

• MCP Infrastructure: Architected 3 custom MCP servers — a PostgreSQL + KeyDB DB inspector (17 tools), a semantic documentation search engine using ChromaDB + sentence-transformers (8 tools), and a zero-overhead API/WebSocket/UDP traffic recorder with ~2,300 writes/sec using WAL-mode SQLite
• GoOracle AI Assistant: Built an intelligent trading operations assistant enabling API-driven trade workflows, real-time price sync, and contextual Q&A powered by modular MCP services
• Smart Order Routing (SOR): Low-latency framework evaluating multi-venue liquidity, spread, slippage, and depth across exchanges using Python + performance-critical C++
• Consolidated Order Book: Real-time Level-2 aggregation engine that merges order book streams from multiple exchanges with sequence alignment and data normalization

Remote | New Delhi, India

• Replaced OpenNMS dependencies with custom service discovery, improving incident detection accuracy by 60%
• Built interactive network topology builder reducing configuration errors by 40%
• Designed real-time performance monitoring dashboards with SNMP integration, cutting incident resolution time by 60%
• Implemented SSH/RDP connection features improving remote network management by 50%
• Built dynamic PDF report generation with Puppeteer (tables, charts, graphs)

LangGraph-based agent automation engine that decomposes high-level tasks into multi-step workflows with dynamic tool execution, memory, and guardrails

Stack: Python · LangGraph · LangChain · FastAPI · Docker

Modular MCP framework enabling AI agents to securely inspect production databases, search exchange documentation, and analyze HTTP/WebSocket/UDP traffic with <0.1ms overhead. Handles ~2,300 async writes/sec across 5+ microservices.

Stack: Python · PostgreSQL · ChromaDB · SQLite · WebSocket · UDP · sentence-transformers

Real-time AI chat app with prompt caching, tool orchestration, streaming responses. Reduced token costs by 30% and improved response time by 40%.

Stack: Next.js · TypeScript · LangChain · LangGraph · GPT-4o-mini · Clerk · Convex

Multi-modal AI platform leveraging OpenAI + Replicate APIs for automated content generation. Boosted user productivity by 50%, reducing manual workload by 30 hrs/week.

Stack: Next.js · TypeScript · OpenAI · Replicate · Prisma · MySQL

M.Sc. Computer Science — MIT World Peace University, Pune (2023–2025) CGPA: 7.5

• 🥇 Certificate of Achievement — Codeinbound LLP (2024) — Led 20% increase in project efficiency
• 🚀 Smart India Hackathon — Advanced to next round with blockchain-based land document digitization platform (MIT-WPU, 2023)

Open to AI/ML Engineer, Backend Engineer, and Full Stack Developer roles in India or globally remote