For the first six seconds
The short version.
Full-stack and platform engineer in Mumbai. I build systems that keep working when the network, the model or the data does not. Graduating 2027, available for full-time SDE, platform, backend and real-time roles.
01 / Position
What I am good at
I work across the browser, the API and the model layer, with a bias for systems that are secure, observable and fast under real use. My strongest work is offline-first sync, AI reliability with explicit fallbacks, and interfaces that make a system's state legible to the person operating it.
- Offline-first architecture. IndexedDB writes, idempotent sync queues, domain conflict rules and a tested reconnect path.
- AI as a dependency. Bounded timeouts, provider fallbacks, confidence gates and audit logs that record which model produced which field.
- Platform and backend. Typed APIs, PostgreSQL schema and access control, Docker deployments and observable failure modes.
- Real-time and auth. WebRTC and Socket.IO signalling, plus session models with HTTP-only cookies, CSRF protection and server-side role checks.
02 / Fit
What I am looking for
Full-time software engineering roles starting now, with a preference for platform, backend, real-time or AI-reliability teams. I care more about the system problems than the job title. Mumbai-based (UTC+5:30); open to discussing the team's working model.
03 / Experience
Where I have worked
Building and maintaining the backend services and platform infrastructure behind the product.
Owned frontend and backend features across React/Next.js and Node/Express. Built an async thumbnail pipeline that cut image load latency by 40%, shipped a reel-generation pipeline, added Sign in with Apple, and closed 10+ production vulnerabilities including IDOR and XSS.
Won the clean and green domain with EcoVision Bharat, a civic waste-reporting platform using YOLOv8 severity analysis, GPS evidence and municipal dispatch workflows.
04 / Selected work
Three records to read
Emergency QR access under 500ms, dual-provider AI fallback under 2s, and SHA-256 document hashes notarised on-chain. The break-glass pass is cached to work without a cellular connection.
Zero data loss on reconnect, a nine-state referral lifecycle, and an OCR confidence gate that never auto-saves an extraction below 90%. Clinicians never wait on the network.
187+ active students and 6.7k tracked events. A transparent 0–100 ATS rubric shows exactly which signals moved the score, across four role dashboards.
05 / Stack
What I build with
Languages
Frontend
Backend
Data
AI / ML
Real-time
Infrastructure
Web3
06 / Questions
The things recruiters ask
What roles are you looking for?
Full-time software engineering roles, available now, with a preference for platform, backend, real-time or AI-reliability-heavy teams. I care more about the system problems than the job title.
Where are you based?
Mumbai, India (UTC+5:30). I am happy to discuss the team's working model, including remote and hybrid arrangements.
What is the strongest evidence of your work?
The three case studies on this site, each with the constraint, the decisions, the architecture and what broke. MediVault and SwasthyaSetu are live, and their source is public. The receipts page links every claim to its evidence.
How do you use AI in your systems?
As a dependency to design around, not a black box to trust. That means explicit fallbacks, bounded timeouts, confidence gates, human review where the stakes are clinical, and logs that record which provider produced which result.
Can I see the code?
Yes, for the public projects linked from each case study. Employer work at Chitralai and Allkane is private; I can talk through the architecture and the measured outcomes on a call.