Applied Research Scientist / Full-Stack Engineer
I build dependable software for the web, mobile, and embedded systems — then poke at it until I find a faster, cleaner, or safer way to do the same thing again. Currently leading applied research at Nuvai AI.
Trajectory
Applied Research Scientist — I
Leading R&D initiatives, coordinating delivery of AI/ML solutions from concept through deployment. Owning research infrastructure: compute, environments, and tooling for experimentation and training at scale. Working across algorithm design and model optimisation.
currently hereCybersecurity Intern
Penetration tested 15+ enterprise applications. Identified and documented 50+ critical vulnerabilities. Mitigated 12 active threats protecting 100,000+ user records. Automated security scanning with Python.
100k+ records securedFull Stack Development Intern
Architected a SaaS platform with Svelte, Tailwind, MongoDB, and Node.js. Built interactive gamification features using Canvas API and WebSocket. Held 95+ Lighthouse scores across all dynamic pages.
10k+ users servedMERN Stack Development Intern
Integrated APM with New Relic and Datadog. Optimised MongoDB queries with a Redis caching layer. Wrote automated Jest + Cypress suites, reducing post-deployment bug reports by 35%.
40% efficiency gainFull Stack Development Intern
Built microservices with ASP.NET Core and MySQL. Delivered pixel-perfect responsive landing pages at 98/100 PageSpeed. Led agile workflow in a 5-developer team with a zero-defect deployment record.
TTFB 1.2s to 720msSelected Work
Production-grade web search engine using Flask microservices and OpenSearch. Distributed caching with Redis and Docker containerization for horizontal scalability. Custom ranking reduced query response time from 250 ms to 150 ms across 50k+ daily active users.
Real-time IoT monitoring at 99.9% uptime across 1,000+ concurrent sensor connections. ML anomaly detection prevented 95% of potential equipment failures; WebSocket + Redux cut data latency by 60%.
Behaviour-cloning CNN trained on 20,000+ hours of simulated driving. Arduino-controlled hardware with a 40 ms control loop. Computer vision for lane detection, obstacle avoidance, and traffic-sign recognition.
Decentralised freelancer marketplace on Ethereum. Escrow smart contracts with milestone payments and dispute resolution — no intermediaries, cross-border by default.
Products
SaaS Platform
A no-code platform to create, customise, and deploy fine-tuned AI models. Visual prompt builder, RAG-based document grounding, OpenAI-compatible API endpoints, and side-by-side model comparison.
Mobile App
A full-featured task app with voice notes, smart deadline notifications, burnout monitoring, Razorpay subscriptions, and Firebase real-time sync. Built for people who need more than a basic to-do list.
Open Source
Open-source libraries I maintain. Built with TypeScript, designed to solve real problems, published on npm.
Multi-layer caching with memory, Redis, and file-store strategies. Auto-invalidation, cache warming, and a real-time analytics dashboard. Adopted by 15+ production apps.
npm i cache-composerAI-powered CLI code reviewer that asks junior-developer-friendly questions. NLP-driven, context-aware suggestions. v2 published on npm.
npm i reverse-stackoverflowModern image pipeline: WebP / AVIF, responsive generation, CDN integration, batch processing for automated workflows. v2 published on npm.
npm i optimix-image-pipelineFramework-agnostic onboarding engine for product tours, interactive tooltips, and progress checklists. Zero runtime dependencies. Adopted by 10+ SaaS apps.
npm i onboard-flow-engineStack
A working set of languages, frameworks, and infrastructure tools — what I use day-to-day and what I've shipped with.
Reach Out
I'm open to full-stack engineering, applied research, and security roles — and happy to chat about side projects, open-source, or interesting problems in general.
Response
I usually reply within 48 hours. Serious inquiries get priority; everything else still gets a reply.