Autonomous Robotics · AI Systems · Embedded Intelligence
I design and build end-to-end intelligent systems — from SLAM-based robot navigation on Jetson Nano hardware to AI-integrated automation pipelines that run on free-tier cloud infrastructure. Systems, not demos.
I'm a third-year EEE student and early-stage researcher at the University of Asia Pacific, Bangladesh. As team leader of a three-member robotics group, I'm building a fully autonomous mobile robot — from hardware integration on NVIDIA Jetson Nano to software orchestration with ROS2, SLAM Toolbox, and Nav2.
My work doesn't stop at the robot. I build AI-integrated automation pipelines that bridge physical systems with cloud intelligence: weekly AI digest systems, academic automation with RAG, and live team monitoring dashboards. I treat free-tier infrastructure as a design constraint, not a limitation — and it makes me build smarter.
I'm actively targeting fully funded MSc or Research Assistant positions in autonomous systems, embedded AI, and robotics for 2027–28.
Full-stack autonomous mobile robot built from scratch, leading a 3-member team. The system implements complete ROS2 navigation — SLAM Toolbox for real-time 2D mapping, Nav2 for global and local path planning, and multi-sensor fusion combining LiDAR and Kinect RGB-D data. Distributed control runs over a Zenoh router with Tailscale mesh VPN, enabling remote monitoring and teleoperation.
AI-powered weekly robotics intelligence digest — now a fully documented flagship. Fetches 40+ RSS feeds across 7 categories, deduplicates, AI-filters with Groq, tracks target-conference CFPs, and narrates a 9-section HTML report with Gemini 2.5 Flash. v1.1 adds offline/demo modes, dry-run artifacts, Docker, and a 35-test suite.
Real-time command center for the Team Supersonic robot build. Live task, handoff, error and note tracking synced via Firebase Realtime Database, PIN-secured member access with server-enforced auth rules, progress charts, and a Groq-powered T2 AI assistant. Single-file static app with full architecture docs.
Multi-model AI pipeline that monitors Google Classroom for new course materials and auto-generates comprehensive engineering master notes using LoRA fine-tuning and Gemini Flash 2.5 — with special handling for EEE topics. Triggered on upload, outputs structured Word documents.
Fully automated AI-integrated LinkedIn management pipeline handling profile optimization, content writing, job discovery, and strategic connection building. Built entirely on free infrastructure: Gemini API, Google Scripts, Google Sheets, and a hosted web app.
AI-powered generalized team project dashboard. Any technical team onboards in minutes — upload your implementation bible, and the AI plans phases, weighted tasks, handoffs and milestones matched to each member's expertise. Works free in the browser, optional Firebase sync.
Generalized, AI-integrated news intelligence automation. Configure any industry — RSS sources, keywords, audience context — and it AI-filters, enriches and emails a personalized digest on a schedule. Fully tested pipeline with graceful fallback.
Private AI career operating system that turns daily engineering work into a funded-MSc/RA application engine: evidence intake, LinkedIn content drafts, professor/lab networking, scholarship intelligence, a first-paper tracker and a daily morning briefing — always human-approved.
Hybrid AI tutor + study-performance analytics + deadline management for EEE undergraduates. Streamlit app with mastery scoring, GPA estimation, weak-topic detection, weekly streaks, Google Calendar prep-block scheduling, and a Gemini 2.5 Flash tutor grounded in your own Google Docs notes. Runs in demo mode until you connect your account.
My all-in-one career command center monorepo: an Express + Firestore backend with a Gemini agent, a web dashboard, an Android companion app (WebView + FCM + offline reminders), a class-routine engine, a Google Classroom connector, and a full restart mission plan (115 tasks, 20 deadlines). Runs entirely on free-tier infrastructure.
Research on autonomous mobile robot navigation integrating ROS2-based SLAM with AI-enhanced decision-making. The system addresses real-time mapping and navigation challenges using SLAM Toolbox, Nav2, and sensor fusion with LiDAR and RGB-D cameras on resource-constrained edge hardware (NVIDIA Jetson Nano). Evaluation focuses on map consistency under sensor noise and navigation performance in dynamic environments.
Target: IEEE-indexed conference (URAI / ROBIO / RAAICON 2026)Real-time 2D mapping with SLAM Toolbox on mobile robot platforms, with focus on loop closure and map consistency under sensor noise.
Deploying vision inference pipelines (TensorRT, YOLOv11) on Jetson Nano for real-time perception within strict compute and power constraints.
Agentic AI systems automating research workflows — from paper discovery and digest generation to academic note synthesis with RAG.
University of Asia Pacific, Dhaka, Bangladesh
Focused on embedded systems, signal processing, and applied AI in engineering contexts. Year 3 with active research running in parallel with coursework.
University of Asia Pacific
Leading a 3-member team on full hardware and software development of an autonomous mobile robot. Responsible for system architecture, research direction, task delegation, and the current paper in preparation.
Regional and International Venues
Competed in IEEE-affiliated regional and international robotics challenges as team leader, including competitions at IEEE Region 10 and other venues.
I'm actively seeking fully funded MSc or Research Assistant positions in robotics, autonomous systems, and embedded AI for 2027–28. If your lab works on any of these, I'd like to connect.