Md. Saidur Rahman Pulok
Hello, I'm Pulok. I’m an engineer who loves to build. While I hold a degree in Mechanical Engineering from BUET, my curiosity led me down a different path: teaching machines to solve complex problems. Today, I work as a Machine Learning Engineer at ACI PLC, where I apply code to automate processes and generate insights. I am particularly passionate about the intersection of AI and real-world impact, focusing on building centralized solutions that drive efficiency. When I'm not training models or refining code, you might find me exploring the latest tech trends, watching movies or lost in music.
Education
Bangladesh University of Engineering and Technology (BUET)
B.Sc. in Mechanical Engineering
News
- Aug 2025Joined ACI PLC as a Machine Learning Engineer.
- Jul 2025Left Brick Line Technology Ltd.
- Mar 2025Joined Brick Line Technology Ltd. as a Software Engineer.
- Nov 2024First ever conference paper accepted at IUT ICCET 2024.
- Jul 2024Graduated from BUET.
Experience
Machine Learning EngineerACI PLC
Architected a distributed, on-premise biometric attendance system utilizing InsightFace and Redis Pub/Sub to process high-throughput RTSP streams with millisecond latency. Engineered a production-grade RAG platform for precision agriculture, orchestrating multi-source data ingestion (Weather, Market Metrics) via FastAPI. Deployed a hybrid ML infrastructure using Docker and MinIO (S3), ensuring high availability across self-hosted and cloud environments.
ACI PLC websiteSoftware Engineer IBrick Line Technology Ltd.
Engineered scalable Python backend services and optimized cloud persistence layers using AWS RDS (PostgreSQL) and DynamoDB. Fortified platform security by implementing Multi-Factor Authentication (2FA) workflows via AWS Cognito and Twilio. Spearheaded the integration of Generative AI features into the flagship product, directly enhancing core user capabilities and product value.
Publications
Machine Learning Approach to Gas Pipeline Leak Detection and Localization
Md Saidur Rahman, Farzana Khandoker, Sadif Sadman Arnov
IUT ICCET, 2024
Projects
InstantID: Distributed Real-Time Biometric System
Engineered a high-throughput enterprise surveillance system capable of processing multiple RTSP streams with zero latency. Implemented a decoupled Producer-Consumer architecture using FastAPI and Redis Streams to handle asynchronous workloads. Achieved sub-100ms inference speeds via InsightFace (ONNX/CUDA) and utilized Qdrant for millisecond-latency vector similarity search.
Omniagro: Agentic RAG Agri-Tech Platform
Architected a scalable multi-platform application delivering real-time agricultural intelligence. Orchestrated a system of six autonomous RAG-powered agents capable of utilizing external tools for weather analysis and yield forecasting. Built a robust backend with FastAPI and LangChain, utilizing Supabase for structured data and Cloudflare R2 for cost-efficient object storage.
Bangla Script Digitization Engine
Developing a deep learning pipeline to transcribe complex handwritten Bangla characters into digital text. Leveraging PyTorch and Hugging Face Transformers to fine-tune Vision Encoder-Decoder models, addressing the challenges of cursive script connectivity and diverse handwriting styles.
And something super cool ;)
You will never guess what I'm working on right now!
GitHub Contributions
Contact
I’m always open to discussing new research opportunities, collaborations, or just a friendly chat about technology and engineering.