Software Engineer & Researcher exploring reliable software systems and AI for software engineering(AI4SE). Based in Dhaka, Bangladesh.
Open to PhD Programs & Research Opportunities
About
I am a CSE graduate with a major in Software Engineering (SE) from AIUB, graduating in 2026. I am currently working as a Software Engineer to understand industry best practices and gain hands-on experience, while also researching how AI is transforming Software Engineering (AI4SE). I have a publication in computer vision and completed my BSc thesis on object detection for autonomous vehicles. My current interests are AI4SE, AI Agents, and LLM reasoning, with a broader interest in AGI. I am open to PhD and research opportunities, but I do not want to limit myself to these areas. I am a learner, and I am open to exploring new fields and interesting problems.
Contributing to the study "Does Messy Mean Wrong? Testing Whether Trajectory Process Indicators Predict Failure Under Strengthened Coding-Agent Evaluation," examining whether process-level signals in a coding agent's execution trajectory predict task failure under a strengthened evaluation protocol.
Contributing to the evaluation methodology to disentangle trajectory "messiness" from outcome correctness in coding-agent benchmarks, addressing limitations of single-turn, outcome-only evaluation frameworks.
CredosisApr 2026 - PresentSoftware Engineer I Remote · Dhaka, Bangladesh
Designing and building the in-house API and booking system with ASP.NET Core Web API.
Built the company frontend with Next.js and TypeScript.
Supervising software engineering interns during product development.
BSc in Computer Science & EngineeringMajor: Software EngineeringThesis: A Two-Stage Object Detection Framework for Autonomous Vehicles Using a Custom Multi-Weather Dataset and YOLOv8CGPA: 3.61 / 4.00
S. Parvin, F. Munsy, M. A. R. Rahat, A. N. Jhumur, K. Nur, D. Ghose
Results in Engineering · 2025
A published study introducing an enhanced YOLOv9 model trained on the Multi-Weather Pothole Detection (MWPD) dataset for robust detection across adverse weather conditions.
Enhancing Factual Consistency in LLMs through RAG and Knowledge Graphs
J. S. Shuvo, M. S. Maahi, M. R. S. Riad, F. Munsy, A. Salam
ICCA 2026
Submitted work on improving the factual consistency of large language model outputs by combining retrieval-augmented generation with knowledge graphs.
Submitted
Does Messy Mean Wrong? Testing Whether Trajectory Process Indicators Predict Failure Under Strengthened Coding-Agent Evaluation
F. Munsy, T. Chowdhury, M. S. Sayed, M. S. Mahmood
In-progress empirical study on whether process signals in a coding agent's trajectory, such as how messy its path looks, predict task failure under a strengthened evaluation.
A LMS built with a focus on security, scalability, and modern design. It features strict 4-tier Role-Based Access Control (RBAC) enforced on the backend, structured course enrollments, persistent progress tracking, server-side graded quizzes, and a fully functional draft-to-publish blog system.
A REST API for a vehicle rental company. Staff log in with a JWT, manage the vehicle fleet, record customer bookings as rentals, and generate monthly revenue reports.
This project is an AI-based job matching and career guidance portal. It uses artificial intelligence to accurately match job seekers with suitable employers and provides personalized career advice to help candidates make informed decisions.