Sheik Sadi Rohmotullah — Backend & GenAI Engineer

I build the systems
AI answers move through.

Backend-first engineer shipping production RAG pipelines, evaluation harnesses, and cloud infrastructure — currently Acting Head of Engineering at a GenAI SaaS company. Based in Dhaka. Building toward Tokyo.

See the systems Get in touch

I'm Sadi — currently Acting Head of Engineering at ELELEM AI, a UK-based GenAI SaaS company, where I own architecture and delivery for two live AI products used by enterprise clients including Sony/SIE. Before taking on that scope I spent two years building the backend myself: RAG pipelines, LLM analytics, and the infrastructure that keeps them running. I care about the unglamorous half of AI engineering — evaluation, quality gates, cost, latency — the work that decides whether a model demo becomes a product people can rely on. I worked a contract role in Tokyo in 2023 and have kept up conversational Japanese since. I'm actively looking to go back.

01 — Selected systems

Things I've shipped to production.

Hybrid RAG Engine

100s of queries/day

A hybrid-search retrieval engine ("Snippets API") with origin-tracked URL redirection, serving live traffic across 8 enterprise clients from client websites in real time.

  • FastAPI
  • ElasticSearch
  • RAG

Centralized Auth (JWKS)

Zero-downtime key rotation

An issuer service minting ES256 JWTs with rotating JWKS, backed by GCP Secret Manager and verified by an in-process caching library across every downstream service.

  • ES256
  • GCP Secret Manager
  • Auth

PII Redaction Pipeline

Client: Sony / SIE

Custom Microsoft Presidio recognizers and a product-specific allow-list scrubbing personally identifiable information from enterprise conversation logs before they reach analytics.

  • Presidio
  • NLP
  • Privacy

Attribution & Analytics

elelem_cid pipeline

A first-party click-ID attribution system connecting widget engagement to backend conversion events, built for enterprise-grade reporting and BigQuery dashboards.

  • BigQuery
  • Cookies
  • Analytics

Recommendation Engine Rebuild

+50% CTA / click-through

Re-architected a recommendation engine on sentence embeddings, validated against the incumbent model through a live A/B test on real enterprise traffic.

  • Embeddings
  • A/B Testing
  • ML

Benchmarking Pipeline

RAGAS · DeepEval · TruLens

An automated report-generation pipeline for evaluating RAG and agentic systems against multiple frameworks, producing ACM-formatted research output.

  • Evaluation
  • RAG
  • Automation

02 — Track record

Where I've worked.

2026 —

Acting Head of Engineering

ELELEM AI · UK, Remote

CTO-level ownership of product and engineering roadmap after the founding CTO's departure. Own architecture, quality, and delivery for two live AI products; lead a 3-engineer team.

2025–26

Lead Engineer

ELELEM AI

Built the Dhaka engineering team from scratch. Owned the full GCP infrastructure stack across two production products.

2024–25

Software Engineer, Backend

ELELEM AI

Built the hybrid-search RAG engine and LLM-based analytics pipelines. Scaled the platform from zero to 8 active enterprise clients.

2023–24

Back-end Developer

ELELEM AI (formerly CONCURED)

Decomposed a monolithic GCP pipeline into independent microservices. Rebuilt the recommendation engine for a 50% lift in CTA/click-through.

2024

Software Engineer, Backend (Contract)

EBLICT — Bangladesh Ministry of ICT

Led backend and ML delivery for a Virtual Private Assistant project across a 10-person cross-functional team.

2023–24

AI Engineer (Contract)

Hiperdyne Corporation — Tokyo, Japan

Built a computer-vision face-privacy pipeline. Researched LLM jailbreaking vulnerabilities to inform secure production deployment.

2022–23

Junior Back-end Developer → NLP Intern

CONCURED

Replaced a paid keyword-extraction service with an in-house open-source solution, cutting a recurring monthly cost.

03 — Stack

What I build with.

AI / GenAI

[ LLM production integration ] [ RAG pipelines ] [ LangChain ] [ LangGraph ] [ OpenAI SDK ] [ Anthropic SDK ] [ RAGAS ] [ DeepEval ] [ TruLens ]

Backend

[ Python ] [ FastAPI ] [ Django ] [ Flask ] [ AsyncIO ]

Cloud & Infra

[ GCP — Cloud Run ] [ GKE ] [ Pub/Sub ] [ Cloud Build ] [ AWS ] [ Docker ] [ Kubernetes ] [ GitHub Actions ]

Data

[ MongoDB ] [ ElasticSearch ] [ Redis ] [ BigQuery ] [ MySQL ] [ PostgreSQL ]

ML

[ TensorFlow ] [ PyTorch ] [ ONNX ] [ Pandas ] [ Scikit-Learn ]