Skip to content
EK

Search

Projects

Case studies, not portfolio cards — each one covers the problem, the constraints, the decisions I made, and what actually changed.

Medical Avenue · 2026

Taking over a React Native app that was OOM-killing itself: 1.2GB → 350MB

Inherited a React Native patient app from a departed agency team, traced a memory profile that climbed past 1.2GB and got the app killed on real devices to the navigation architecture itself, and rebuilt it to a stable ~350MB — verified on a 2018 Android with 3GB of RAM.

  • 1.2GB → ~350MB stable
  • OOM kills on devices → zero
  • Verified on 2018 Android, 3GB RAM
Read case study →

Medical Avenue · Dec 2025 – Jan 2026

A RAG system built in 2.5 weeks to walk into a CEO meeting with proof

Built a source-grounded RAG assistant for the medical-tourism domain — solo, before any contract existed. An ingestion pipeline over 264 interview videos, 2 books, and 7 papers; pgvector retrieval with MMR; and answers that cite their sources or don't answer at all.

  • 264 videos + 2 books + 7 papers ingested
  • Cited, source-grounded answers only
  • ~2.5 weeks, solo, pre-contract
Read case study →

NayaOne · 2025

FCA–NVIDIA AI Sandbox: a launch feature shipped in one week

Built the product module behind the FCA–NVIDIA AI Sandbox launch in one week — full-stack, shipped dark behind feature flags on a live platform serving 80+ institutions, switched on for a fixed public launch date.

  • 1-week delivery deadline
  • Launched on schedule — June 2025
  • Official FCA technology partner
Read case study →

Want to talk about a role or a system like these?