I'm a Senior Software Engineer at SELISE Group in Dhaka, Bangladesh. I specialize in building secure, high-concurrency systems, generative AI ecosystems, and distributed architectures.
My work spans across Backend, Mobile, and Web technologies, with a focus on IoT, RAG pipelines, and Multi-Agent Systems. I previously worked on creating scalable solutions for Swiss clients in sectors like construction, sports, and smart metering. I hold a Bachelor of Science in Computer Science & Engineering from Gopalganj Science and Technology University (GSTU).
Email / CV / LinkedIn / Scholar / Github / Kaggle / Medium / ResearchGate
Academic research focused on Computer Vision, Natural Language Processing (NLP), and Deep Learning. Published works demonstrate expertise in developing novel algorithms, achieving high accuracy in classification tasks, and applying AI to real-world biological and medical problems.
Achieved 100% classification accuracy on local fish datasets by developing a novel methodology that integrates U2-Net saliency detection with Histogram of Oriented Gradients (HOG) and ensemble classifiers.
Enabled automated fish cultivation monitoring by prototyping a comprehensive census system (FOCS), utilizing deep learning for species classification and IoT sensors for real-time environmental data acquisition to optimize aquaculture efficiency.
Achieved 96.67% diagnostic accuracy in detecting Parkinson's disease by engineering an image-based classification tool that utilizes fine-tuned DenseNet121 and InceptionV3 architectures to analyze hand-drawn spiral and wave patterns.
Enhanced local language processing capabilities by authoring `Bangla`, an open-source Python library that implements efficient stemmers and lemmatizers using optimized Trie data structures.
Professional and personal software engineering projects showcasing scalable architecture design, full-stack development, and AI integration. Proven track record of delivering high-performance mobile and web applications, microservices, and enterprise solutions using modern technologies like Flutter, and Python.
Engineered a secure, on-premise Enterprise Generative AI Ecosystem, reducing model inference latency by ~15% and driving daily active usage, by architecting a high-concurrency LangGraph orchestrator with Redis-based distributed cancellation, integrating Whisper/Perplexica agents, and implementing a real-time collaborative chat engine with secure invitation tokens.
Seamlessly extended the Enterprise AI Ecosystem to mobile platforms, increasing platform engagement by 40%, by engineering a high-performance feature-parity Flutter application with real-time WebSocket chat and offline-first media capabilities.
Unified the Enterprise AI experience for thousands of users by architecting a high-performance React/Zustand web platform with a Unified WebSocket Protocol, delivering 100% feature parity and a 60fps rich-text chat interface.
Optimized ad spend and targeting efficiency by architecting an intelligent marketing engine that leverages Python and Machine Learning to autonomously interface with multiple platforms and deploy data-driven ad campaigns.
Delivered an all-in-one fan engagement platform for over 30k users by architecting a modular Flutter application (`tickets`, `race-center`) that integrates Bitmovin for sub-3s latency live streaming and real-time telemetry from FIS data providers.
Engineered a scalable, industry-agnostic IoT middleware orchestrating millions of telemetry messages for Maritime, Smart Metering, and Industrial use cases, featuring a Netty-based transport layer and a drag-and-drop Rule Engine.
Solved daily commute challenges for university students by designing a real-time vehicle tracking application, integrating a Bengali-language chatbot and Geolocation services for accurate live updates.
Enhanced emergency response efficiency by creating a comprehensive fire safety ecosystem, combining IoT hardware sensors with a mobile app to provide real-time alerts and precise victim location tracking.
Designed a communication-based satellite simulation capable of atmospheric sounding and controlled descent, achieving an Honorable Mention in the CanSat Competition 2019 by engineering the bi-directional telemetry system and Ground Control Station.
Engineered a remote-controlled submersible surveillance vehicle featuring a custom PVC chassis and waterproof propulsion system. Designed the electronic control unit using Arduino and RF modules to ensure reliable underwater maneuverability and signal transmission.
A collection of my academic teaching experience, professional recognition, and community service. Includes roles as a lecturer, competitive programming achievements, and leadership in technical clubs.
Served as a part-time lecturer for the Computer Technology Department and managed institutional websites.