Software Engineer (AI)
FINTECH COMPANYPromoted after 5 months. LLM workflows behind Bangla-language reports, corporate documents & video content; pipelines and dashboards turning 350K+ transaction records into numbers field managers use.
I'm Rimon Debnath — I design and engineer full-stack products and LLM-powered systems that ship to real users, from fintech dashboards to production AI workflows.
I turn ambitious ideas into production software — currently building the AI systems behind a cardless-EMI smartphone financing platform, where my LLM workflows turn days of manual work into minutes.
Promoted from Frontend Developer to Software Engineer (AI) in five months at a fintech company. Before that I built high-availability AWS infrastructure for client projects at Neovatech and shipped Node.js REST APIs with Firebase real-time features at Younglabs. I obsess over clean architecture, honest timelines, and interfaces that feel expensive.
Next.js · React · TypeScript · Node — design to deploy.
LLM workflows, document & report engines, data pipelines.
AWS architecture, Docker, CI/CD — 99.9% uptime on Neovatech client infrastructure.
Promoted after 5 months. LLM workflows behind Bangla-language reports, corporate documents & video content; pipelines and dashboards turning 350K+ transaction records into numbers field managers use.
Production features in Next.js, React & Tailwind for the EMI platform, working directly with the founders on product direction; debugged production faults, supported backend/API integration and handled database operations including security and scheduled backups.
High-availability AWS at 99.9% uptime through 200% traffic spikes; microservices and APIs across 7+ client projects for 10K+ monthly users, with −40% latency; Docker + GitHub Actions CI/CD cut releases from 3 days to under 24 hours.
Node.js REST APIs for 500+ concurrent users at 99.5% uptime; MVP shipped in 12 weeks with Firebase Auth and Realtime Database.
Education. Bachelor of Technology in Computer Science and Engineering, Jalandhar, India.
Three live side projects with public code, and three systems from my day job at a fintech company, whose code is private.
DRAG, SCROLL SIDEWAYS OR USE ← →SWIPE TO EXPLORE →
LIVE
Multi-tenant security portal: turns raw Nessus and Nmap scans into executive reports. Findings are sanitised before storage, a local LLM drafts the summary, and access is gated by National ID checks.
LIVE
Open-source, Linear-style issue tracker: live multi-user boards on Supabase Realtime, row-level security on every table, optimistic updates with rollback.
CONCEPT
A redesign concept for a Dhaka gym, built on my own initiative: parallax hero, scroll-driven services gallery, stacking process cards, fully responsive and reduced-motion aware.
UNOFFICIAL CONCEPT, NOT CLIENT WORK
LLM workflows behind Bangla-language operational reports and corporate documents: jobs that took days of manual work now finish in minutes.
Production frontend for a cardless-EMI smartphone financing platform used by field operations across 7 divisions of Bangladesh.
LLM workflows behind Bangla-language video content, part of the AI systems I build in my day job.
HAVE AN IDEA IN MIND?
LET'S SHIP SOMETHING RARE