Aperture
Rebuilding a fintech dashboard — cutting time-to-insight from minutes to seconds.
One canvas layer renders the entire candlestick series; a WebGL pass holds 1M+ points at 60fps while a lightweight D3 overlay keeps axes razor-crisp.
Enterprise SaaS backend systems, REST APIs, and AI features — engineered to stay fast, correct, and unremarkable under real load.
Not a gallery of mockups. Each project shipped to production, moved a metric that mattered, and paid down long-term quality.
Rebuilding a fintech dashboard — cutting time-to-insight from minutes to seconds.
One canvas layer renders the entire candlestick series; a WebGL pass holds 1M+ points at 60fps while a lightweight D3 overlay keeps axes razor-crisp.
I start from the data model, not the library. When a query is slow I read the query plan before touching the code.
Production issues are feature requests. I read the logs first, then refactor so the next engineer doesn't trip.
The metric, not the ticket, is the requirement. A feature that ships on time and stays reliable wins.
LLMs are only useful grounded and evaluated. Retrieval, not prompts, is where the quality lives.
Clean schemas and fast queries beat clever code. The best backend is the one nobody has to think about.
I profile against real load before calling it done — query plans, latency, and the p95, not the happy path.
Schema and contracts before code. Interfaces are the product — everything else is implementation detail.
Tests, reviews, and refactors are the norm, not the exception. The code you inherit is code you improve.
Perfect is the enemy of shipped. Get it to production, watch the metrics, and iterate on reality.
B.Tech in AI & ML — learned models are easy and production is hard.
Grounded LLMs on real company knowledge, not a demo corpus.
Prompt engineering, evaluation, and shipping AI features to enterprise customers.
Agentic helpers for my own workflow — to see what actually holds.
Production ML applications that quietly do a job end-to-end.
Peoplebox.aiRemote
Build production-grade backend services and AI-powered features for enterprise SaaS.
Peoplebox.aiRemote
Learned the enterprise SaaS stack on the job while shipping real, customer-facing work.
Let's work together
I'm looking for teams that value scalable backend engineering, well-designed APIs, and AI built on solid fundamentals. If that's you, say hi.