12
AI projects delivered
3
Peer-reviewed papers
8
Technical domains
2024
Building AI since
Md. Sakibur Rahman is an AI developer at Sparktech Agency in Dhaka, Bangladesh, building retrieval-augmented and agentic AI systems — the kind that reason, retrieve, and act, rather than only generating text.
He joined Sparktech as a trainee in October 2025 and moved through junior to his current role, working across RAG and agentic RAG pipelines, LLM reasoning grounded in real retrieval, NLP components, and AI chat assistants, carrying features from prototype through to production. The emphasis throughout is on reliability and explainability: grounding model outputs, testing edge cases, and understanding the retrieval and reasoning layers beneath the interface.
Alongside product work he publishes applied AI research, with peer-reviewed contributions spanning medical image diagnostics, source-code understanding, and brain-computer interfaces, presented at international conferences including ICDMIS 2025 and IEEE SPICSCON 2025.
AI & retrieval
Language & generation
Vision & signals
Engineering
Three promotions in under a year at Sparktech Agency, each one widening the scope from learning production RAG to owning it.
July 2026 — Present
2 mosNowSparktech AgencyDhaka, Bangladesh (On-site)
Step 04 / 04
January 2026 — June 2026
6 mosSparktech AgencyDhaka, Bangladesh (On-site)
Step 03 / 04
October 2025 — December 2025
3 mosSparktech AgencyDhaka, Bangladesh (On-site)
Step 02 / 04
November 2024 — May 2025
7 mosAcote GroupDhaka, Bangladesh (Hybrid)
Step 01 / 04
Reliable, explainable, production-grade AI: grounding model outputs, testing edge cases, and understanding the retrieval and reasoning layers beneath the interface.
01
Retrieval-augmented and agentic pipelines that ground model output in real sources — so an answer can be traced back rather than taken on trust.
02
Designing production LLM systems — prompting strategy, structured outputs, and provider-agnostic abstractions that swap models without rewriting a product.
03
Working with language models beyond generation — understanding, classification, summarization, and the preprocessing that makes them viable.
04
Applying generative models to product surfaces where the output has to hold up in front of a user.
05
Building and evaluating image classification systems, including for diagnostic use cases.
06
Real-time EEG signal processing for assistive control — the subject of his published BCI research.
07
Adapting models to a task and measuring the result honestly against a baseline — including the unglamorous data work that decides whether a model is trustworthy.
08
Carrying an AI initiative from a research question or product idea through to production.
12 projects across LLM products, retrieval systems, computer vision, and speech. The deck cycles automatically — drag, swipe, or use the arrow keys.
A live Qur'an app with Qalam, an AI assistant whose answers are grounded in Qur'an and Hadith retrieval.
01 / 06 — Quranity
3 peer-reviewed publications across medical image diagnostics, source-code understanding, and brain–computer interfaces.
01
2nd International Conference on Data Mining and Information Security (ICDMIS 2025) · Springer
02
4th IEEE SPICSCON 2025 · IEEE
03
3rd IEEE Robotics, Automation, and AI Conference (RAAICON 2024) · IEEE
October 2021 — September 2025
Bachelor of Science in Computer Science and Engineering · GPA 3.56 / 4.00
Artificial IntelligenceMachine LearningImage Processing & Brain–Computer InterfacesBlockchain in Green ITComputer Networks & SecurityAlgorithms & Data StructuresSoftware EngineeringDBMS
Open to AI engineering roles, applied ML work, research collaborations, and AI product consulting — based in Dhaka, working with teams anywhere.