Headline

🏅 Gold Medalist (IIT Goa) | AI/ML × GenAI | Building Scalable Production AI Systems

About

AI/ML & GenAI Engineer focused on building production-grade intelligent systems. I work across the ML lifecycle—from model development and evaluation to deployment, monitoring, and optimization—with hands-on experience in classical ML, LLMs, RAG, AI agents, MLOps, and cloud infrastructure. My recent work includes building LLM-powered agents using LangGraph and Azure OpenAI, production AI/ML systems with FastAPI and MLflow, model monitoring and drift detection, and scalable deployments using Docker, Kubernetes, AWS and Azure. I also have a strong engineering background in distributed systems and Web3/ZK, which gives me a broader perspective on building scalable and reliable products. Focus: AI/ML • GenAI • LLMs • AI Agents • MLOps • ML Systems • Cloud • Production AI

Location

Bengaluru

AI stack

langchain

langgraph

Work experience

Software Engineer at Optum, Aug 2025 - Present · 1 yr 2 mos, Gurugram, 1. Al-Powered Feedback & Session Analysis Engine: Operational Excellence (OE) Agent Verbatim & Member’s Session Analysis Platform, Architected LLM-driven agent using LangGraph and Azure OpenAl to analyze 800+ monthly customer verbatim feedback entries with < 3s latency across provider search and cost price applications. Integrated Session Analysis module with Datadog Session Replay for pattern detection, sentiment analysis, and behavior anomalies. Built automatic Datadog query generation for 2. Problem Analysis with root-cause investigation and stats-driven recommendations. 3. Worked a semi-supervised NLP topic modeling (NMF/LDA) with TF-IDF, NLTK preprocessing, tuning, on 5000+ survey responses to automate theme discovery and classification of verbatims. 4. AI Driven Daily Monitoring Report and Automation of Datadog Notification Alerts over teams/email. 5. Full-Stack Platform Development: Built production FastAPI backend (Python 3.11) with JWT authentication, async request handling, and role-based access control and coordinate with DevOps team for end to end deployment. 6. User Story, Defect & Incident Resolution: Developed and delivered user stories, resolved defects, conducted code reviews, collaborated with cross-functional teams to maintain high code quality, and performed root cause analysis of API failures and production incidents to implement timely resolutions. 7. Received Bravo from Director Software Engineering (Pronil Sengupta) of provider search experience and presented over Learning Coalition Event, 800+ attendants. 8. Technologies Used: FastAP| | LangGraph | Azure OpenAl | React 18 | GenAI | RAG | AWS (EKS, ECS, ECR, Secrets Manager) |Kubernetes | Docker | GitHub Actions | Datadog | Python 9. Notable Metrics: 3s avg latency | 99.9% uptime SLA | Auto-generated Datadog queries | Real-time session insights 10. Currently working on an AI-powered Datadog Monitoring Agent with Datadog MCP Server.

Full Stack Engineer at Tevaera, Jun 2024 - Apr 2025 · 11 mos, • Token Launch & Governance: Led architecture for a cross-chain Smart Contracts with upgradeable/non upgradeable contracts, DAO integration via Tally, and layerZero- powered ERC20 Token. • NFT & Renting Marketplace: Built NFT marketplace and Rental Marketplace with escrow-free ERC-4907 rentals, dynamic fee logic, and seamless cross-chain support. • Teva DEX: Deployed a uniswap V2-inspired AMM with whitelisted LPs and in-game currency swaps. • Guild Protocol: Build GameFi on a bonding curve-based MemeFi and AgentFi protocol layer (inspired by Pump.fun & Virtuals) with NFT-bound Guild ownership, veToken staking, SubDAO governance, programmable liquidity routing, and platform fee streaming via modular gas-optimised smart contracts. • Audits & Security: Audited all major modules using Slither and Testing with Foundry—ensuring gas efficiency and protocol safety. Audited more than 35K lines of code and 80+ smart contracts across the entire Tevaera ecosystem. • SDK and Backend – Delivered SDKs, subgraphs, for seamless frontend integration across all modules, Lead Team of frontend and backend developers for smooth integration and logical correctness. Contributed to Frontend and backend fulfil gap on logic side. • Economic Security Simulated adversarial token/DEX interactions to ensure MEV resistance, incentive alignment, and protection against economic exploits and value leakage. • Technologies Used: Solidity, Foundry, Slither, ERC-4907, Governance Standards, Uniswap AMM, EIP-2535, Token Bound Accounts, Clone Factory, Staking , Typescript, Python, SubDAO, Auditing, Cross Chain LayerZero

Zero Knowledgable Proof(ZKP) Researcher at Tevaera, Jan 2024 - Jun 2024 · 6 mos, Canada, Built an on-chain ZK-powered Social Deduction Game (MVP) using Circom and zk-SNARKs for private role logic—demoed at EDCON 2025. Optimized gas via off-chain proof generation (Gevulot, Risc0) and efficient on-chain verification. Researched zkSync’s prover stack and designed a threshold-cryptography-based session system for secure, social login-compatible Account Abstraction.

Product Lead at The Social Continent, Sep 2023 - Jun 2024 · 10 mos, Built SoCon, a decentralized social identity layer with portable profiles, social recovery, and data ownership—bridging Farcaster & ATProtocol principles. Architecture Protocol (Foundry), P2P node infra, and real-time engagement features. Integrated AI-based interest matching using LLM and scaled via AWS. Designed tokenomics to align incentives for user-owned identity and social graphs. Authored the whitepaper for The Social Continent—A Sufficiently Decentralized solution designed to onboard the next million users.

Blockchain Developer at The Social Continent, Jul 2023 - Sep 2023 · 3 mos, Design a smart contract for a complete decentralized social layer that provides users with on-chain identity and enables on-chain verification of off-chain data, establishing a sufficiently decentralized architecture where users are the true owners of their data. Inspired by farcaster and atProtocol (Fund Raises $180M +)

Blockchain Developer at Cognam Technologies Pvt. Ltd., May 2023 - Jun 2023 · 2 mos, New Delhi, Delhi, India, Built a robust backend to support DeFi operations and provide an integrated card solution for blockchain assets, card settlements are handled through a decentralized node with P2P sync. Technology Involved Spring Boot (Java)

Product Team Lead at Dextr, May 2022 - May 2023 · 1 yr 1 mo, Delhi, India, Engineered ( Proof of Concept) a modular and privacy-enhanced DEX protocol focused on user custody, seamless composability, and smart account UX. Integrated zk-SNARK-based Multi-Factor Authentication for secure, private fund access, and Social Recovery for self-custodial key management. Designed ERC20 Vaults and Portfolio ABTs (SBT-powered identity tokens) to unify user asset views across pools and strategies. Enabled gas-optimized vault deployment via EIP-1167 clone factories, and introduced One-Sided Liquidity Provisioning for flexible LP participation without impermanent loss risks. Leveraged Account Abstraction to simplify on-chain interactions and power advanced trading flows through smart accounts. Built withdrawal and interaction logic in Circom for fully private exits.

Blockchain Developer at Dextr, Jan 2021 - May 2022 · 1 yr 5 mos, New Delhi, Delhi, India, Designed a gasless quadratic voting platform to drive fair, sybil-resistant decision-making in DAOs. Enabled ERC20-based QV tokens, customizable SBTs for identity binding, and ZK-SNARK integration for anonymous surveys. Streamlined UX using meta-transactions and IPFS for decentralized storage.

Education

Bachelor of Technology - BTech, Mathematics and Computer Science, Indian Institute of Technology (IIT) Goa