Hi, I'm Shafi.
Senior Software Engineer · SaaS · Platform Engineering · Distributed Systems
Backend and platform engineer with about ten years building SaaS. I design systems that remove bottlenecks — reporting, data ingestion, reliability — and the libraries and tooling that make other engineers faster. Based in Bangladesh, working globally.
Open to senior/staff/lead engineering, platform and technical-lead roles — happy to talk founding-engineer journeys

Impact in numbers
as of October 2026- Days → self-serve
- report delivery — customers generate their own, no engineering queue
- ½ day → 2 min
- template maintenance, via automation
- 20+ GB
- largest report, streamed straight to S3
- ~2–2.5M
- data-ingestion events per day
- ~60%
- fewer event-processing failures
Proof — read one, or read all four
All case studies →Team & project recovery
Rebuilt a demoralized team and delivered the projects that had already missed their deadlines.
Multiple
missed projects delivered
Self-service reporting platform
Customers generate their own reports — no engineering queue, and template changes ship in about two minutes.
Days → self-serve
report delivery — no engineering queue
DataLake & the type-safe ingestion library
One library turned data ingestion into a minimal-effort integration for every backend team.
~2–2.5M
ingestion events per day
Event & webhook reliability
Queue fairness, retries and exponential backoff — generalized so the whole platform fails less.
~60%
fewer event-processing failures
100/100 × 4 Lighthouse
Met the COO's challenge at Japan Travel with custom tree-shaking tooling — this site scores the same.
4×100
Performance · Accessibility · Best Practices · SEO
Shipped
- MyDokan — my own product, live and distributed privately. Email me for access.
- loop — discussion rooms with LLM personas and a deterministic pipeline to run the work. Open source.
- portfolio MCP — persistent project memory for AI agents. Open source, 27 tools.
Now
Senior Software Engineer at TestReach, building reporting, data-ingestion and reliability infrastructure for a global assessment platform — and building with AI agents daily.
Open to senior/staff/lead engineering, platform and technical-lead roles — happy to talk founding-engineer journeys
Skills
The honest depth map →Daily and deep: TypeScript, Node.js, NestJS, PostgreSQL, MongoDB, Redis, Jest, AWS (event-driven, microservices). Strong regular use runs from Kubernetes, Elasticsearch and Docker to Next.js, Flutter/Dart and Supabase — grouped by honest depth with where each was used on the About page, not percentage bars.
Latest writing
All posts →Fix Angular compatibility issue
Fix “the library which declares XModule is not compatible with Angular Ivy” with the Angular Compatibility Compiler (ngcc).
Optimizing API Response Times - How We Took It from 7.5 Minutes to Under 3 Seconds
How small changes to our MongoDB queries slashed API response times by over 99% — from 7.5 minutes to under 3 seconds.
Add Duck Theory
Where 'adding a duck' comes from — a 90s game-development story about change requests.