status: operational · Dhaka, Bangladesh 🇩🇰
I'm not here looking for a job — I've got one. This exists because there's no clean way to hand someone my actual range in a single link. I started on a PL/SQL pipeline nobody else wanted to touch, and ended up owning backend, frontend, infrastructure, databases, security, and the AI layered on top of all of it. If something needs building or fixing, email is the fastest way to reach me.
Six-to-ten service platforms, zero-downtime deploys
PostgreSQL · Oracle · ClickHouse
K8s cluster, HA load balancers, full CI/CD
RAG, vector search, multi-model routing
Some of this runs under government confidentiality, so names and specifics are generalized — the architecture and my role in it are real.
Replaced manual, leakage-prone paper delivery with a time-gated, cryptographically scoped distribution platform for exam centers nationwide, plus a tamper-evident result archive with per-row integrity checks going back two decades.
Ten-service platform for national food distribution — dealer registration, beneficiary tracking, cross-ministry reporting. Designed and run the Kubernetes cluster and CI/CD end to end, with no single point of failure at the entry layer.
An observability platform that sits in the query path instead of beside it. A Go-based sniffer decodes database wire protocol in real time — Go over the JVM to kill the GC-jitter that drops packets at high query rates — streamed through Kafka into ClickHouse, with a proxy firewall that blocks policy-violating queries before they execute.
Upload scattered documents, describe the report you need — a multi-turn agent retrieves the right chunks via vector search and synthesizes a structured output. RAG over fine-tuning, because the source documents change per user.
[photo]
I didn't set out to become the person who gets called when a database is on fire. That happened by accumulation. First it was a PL/SQL pipeline on legacy Oracle infrastructure that nobody else wanted to touch. Then a six-service monitoring backend. Then a Kubernetes cluster running a platform that a national ministry depends on staying up. Then an AI pipeline, because by that point owning "the layer nobody else wants" had become the job description. Each one taught me the layer below it wasn't optional to understand.
Since May 2024 I've been at IICT, BUET, embedded on government-scale assignments where downtime isn't a metric, it's an incident. I started as Assistant Programmer and was promoted to Programmer for the same reason I keep getting handed the harder systems: I make the architectural call — ClickHouse for time-series throughput, Go over the JVM to kill GC-jitter, RAG over fine-tuning because the documents are dynamic — and I take it through to production myself.
I'm not on here looking for a job — I already have one. This site exists because there's no good way to hand someone my actual technical range in a single link. If you need something built or fixed — backend, frontend, infrastructure, AI, a database that needs saving — email me and tell me what it is.
ruhan.9mri@gmail.com →What I actually build with, end to end.
Where I've worked and studied.
IICT, BUET (promoted from Assistant Programmer)
May 2024 — PresentMerilsoft (Remote, Part-Time)
Jan 2025 — Oct 2025Khulna University of Engineering & Technology
2018 — 2024Notre Dame College, Dhaka
2018No job hunt here — just email. Tell me what needs building, fixing, or rescuing, and I'll tell you if I'm the right person for it.