Headline

AI Engineer & Tech Lead | I build AI products, agents & intelligent systems | Python · Full-Stack · Computer Vision

About

I build AI systems end to end. For me, engineering is a craft - part problem solving, part architecture, and part knowing how all the pieces should come together. I've spent the last nine years building software across frontend, backend, AI and computer vision. Today, I lead a 15-person engineering team while staying close to the code, architecture and the realities of getting systems into production. I enjoy difficult problems - especially the ones that don't have an obvious answer. I like understanding things deeply, getting into the details, and turning ideas into systems that work in the real world. I also care deeply about the people I work with. I enjoy mentoring engineers, helping them think through problems and find their own way of building well. Over time, I'd like to contribute more broadly to the engineering community by sharing what I learn and helping others grow. Outside of work, I build things that interest me - from open-source AI code review and developer tooling to experiments in agent orchestration and agentic infrastructure.

Location

Greater Delhi Area

AI stack

claude code

codex

cursor

deepseek

hermes agent

hugging face

Work experience

Tech Lead at InfraX AI, Nov 2025 - Present · 1 yr, I lead 15 engineers across AI, backend, frontend and edge systems. I set technical direction, write and review code, interview engineers and work with the team through client deployments and production incidents. • Built traffic-operations agents for incident triage and device queries using Python, LangGraph and MCP. Added role-aware tools and confirmations for sensitive actions, plus contract tests across Python and Node.js to catch missing tools and required inputs. The control room these support handles around 200 incidents a day. • Made AI evidence review inspectable: the pipeline samples video frames, extracts structured findings and applies deterministic checks. Operator traces and shadow mode let the team examine decisions before enabling enforcement. • Architected VIDES video analytics deployed on 40 cameras across 10 gantries, combining tracking, multi-frame plate OCR and radar/camera integration. Trained models in PyTorch and exported the detector to ONNX for CPU inference. • Removed blocking waits for incident footage so camera workers can keep processing. Added staged releases, rollback and HMAC-signed webhooks with retries and dead-letter redelivery. • Built the VMS for a 10 km highway corridor: ONVIF camera discovery, PTZ controls, WebRTC/HLS live views and synchronized playback. Added retries and crash recovery so pending recording uploads can resume after failures. • Delivered fare-rule administration and vehicle-exit workflows using number-plate recognition for a live parking site. • Built Steward, an agentic OS prototype for infrastructure, with approved commands and recovery checks against sensor readings. Demonstrated it with simulated building equipment.

Team Lead & Senior Software Engineer at CODEYUG WEB SERVICES PRIVATE LIMITED, May 2023 - Oct 2025 · 2 yrs 6 mos, I led eight engineers across frontend, backend and test automation on AssessPrep, taking features from design through release. I started building AssessPrep's AI capabilities in 2022 and continued that work after my promotion to Team Lead & Senior Software Engineer in May 2023. Selected work across my time on AssessPrep: • Built AI authoring across frontend and backend so teachers could generate questions, explanations and mark schemes from instructions, PDFs and images. • Implemented AI-assisted grading and feedback for typed and handwritten answers, with teacher review controls and automatic checks for questions the platform can verify. • Developed web and mobile features with React, React Native, AngularJS and Node.js, and built automated tests to protect releases. AssessPrep served more than 800 schools by the time I left. My responsibilities grew from frontend development to full-stack engineering and team leadership as Codeyug grew from around 10 to 50 people.

Full Stack Developer at CODEYUG WEB SERVICES PRIVATE LIMITED, May 2021 - Apr 2023 · 2 yrs, I expanded into backend development after three years building the AssessPrep frontend. • Developed Node.js services alongside web interfaces in React and AngularJS. • Started building AssessPrep's AI capabilities in 2022, handling both frontend and backend implementation. • Worked on AssessPrep's companion React Native app. I moved into the Team Lead & Senior Software Engineer role in May 2023.

Frontend Developer at CODEYUG WEB SERVICES PRIVATE LIMITED, May 2018 - Apr 2021 · 3 yrs, I built AssessPrep's assessment and learning interfaces with JavaScript, React and AngularJS. After three years on the frontend, I moved into full-stack development and took on backend responsibilities alongside interface work.

Frontend Developer at PandoCorp Private Limited, Sep 2017 - May 2018 · 9 mos, Bangalore, India, I moved from intern to full-time Frontend Developer after one month, then became the enterprise product's sole frontend developer. I built and maintained AngularJS interfaces for logistics master-data management and Pando Optima's route and cost optimization workflows.

Frontend Developer Intern at PandoCorp Private Limited, Aug 2017 - Sep 2017 · 2 mos, I built logistics interfaces for master-data management and Pando Optima's route and cost optimization workflows. The internship was planned for two months. After the first month, I moved into a full-time Frontend Developer role based on my work.

Internship at Scaledesk Web Studio Pvt. Ltd., Jun 2017 - Aug 2017 · 3 mos, Greater Delhi Area, I built the company's website UI and a landing page called Scala during my web development internship, using HTML, CSS, Bootstrap, JavaScript and jQuery.

Education

Bachelor of Commerce (B.Com.), Commerce, School of Open Learning (DU SOL), University of Delhi

Computer Software Engineering, Navgurukul