Md. Salman Hossan Prottoy
About Me
I am a Master's student in Artificial Intelligence at the University of Jyväskylä, Finland, and a software engineer focused on building scalable, production-ready systems. My interests span language models, retrieval-augmented generation, full-stack development, and cloud-native architectures.
Alongside industry work, I enjoy working on applied research problems at the intersection of natural language processing and trustworthy AI, with prior experience on fake news detection for Bangla.
Experience
Software Engineer (Remote)
September 2023 — Present- Architected and delivered scalable, end-to-end full-stack solutions for multiple major projects, including payment integration and automated communication services.
- Engineered core components for a Retrieval-Augmented Generation (RAG) application, integrating LLMs into a containerized, cloud-native architecture.
- Designed and implemented CI/CD pipelines to automate end-to-end deployment.
- Mentored teammates and established engineering best practices.
- Designed and developed scalable APIs and backend features.
- Maintained and optimized cron jobs and migration scripts.
- Integrated Elastic File System and improved platform UX where needed.
Junior Software Engineer
November 2021 — August 2023- Developed and maintained scalable RESTful APIs using Django, then migrated and optimized them to Go, implementing best practices for authentication, authorization, and data validation, supporting 5,000+ daily active users.
- Engineered and optimized critical data processing pipelines using cron jobs with parallel processing capabilities, reducing processing time by 70% and improving system reliability for enterprise-level applications.
- Designed and implemented comprehensive database migration strategies, integrating AWS Elastic File System for distributed storage solutions, enabling seamless data transitions across multiple database platforms while improving file access performance by 45%.
- Enhanced enterprise application user experience by developing reusable Vue.js components and implementing responsive design patterns, resulting in 30% improvement in user engagement metrics and streamlined interface workflows.
Undergraduate Teaching Assistant
May 2021 — August 2021- Provided academic support for the Web Technologies course (≈80 students).
- Assisted in lesson planning and preparation of instructional materials.
- Collaborated with faculty to identify learning gaps and implement targeted interventions.
Education
Master of Science — Artificial Intelligence
University of Jyväskylä, Finland
September 2025 — Present
Bachelor of Science — Computer Science & Engineering
American International University-Bangladesh
January 2018 — September 2021
HSC — Science
Chattogram Cantonment Public College
2017
SSC — Science
Chattogram Cantonment Public College
2015
Research
Undergraduate Research Assistant
2021American International University — Bangladesh
- Conducted background studies on fake news detection and imbalance handling.
- Collected and analyzed datasets; co-authored a paper presented at IC4IR 2021.
Publications
Approaches for Improving the Performance of Fake News Detection in Bangla: Imbalance Handling and Model Stacking
Proceedings of International Conference on Fourth Industrial Revolution and Beyond 2021
Lecture Notes in Networks and Systems, vol 437. Springer, Singapore
Authors: Hossain, M.M., Awosaf, Z., Prottoy, M.S.H., Alvy, A.S.M., Morol, M.K.
Honors & Awards
Vice Chancellor's Award
March 2023American International University-Bangladesh
Vice Chancellor's Award for Meritorious Effort in the completion of the design project entitled "Approaches for Improving the Performance of Fake News Detection in Bangla: Imbalance Handling and Model Stacking."
Extra-Curricular
Cadet Corporal
2015 — 2017Bangladesh National Cadet Corps (BNCC)
- Completed BTE-01/2016-17 training.
- Trained junior cadets and assisted with event organization.
Projects
A collection of projects I've built across full-stack development, AI research, and cloud infrastructure.
Forest Fire MAS — BDI Multi-Agent System
Autonomous forest fire emergency response built with JADE (Java) and extended with a Python BDI layer using CrewAI and Ollama. Four agents cooperate across fire detection, severity assessment, resource coordination, and emergency dispatch — with multi-zone conflict resolution, uncertainty handling, and belief revision in cascading scenarios.
UniPilot — AI Academic Copilot
AI-powered academic workspace built on Finland's shared higher education data infrastructure (HEDS / Opin.fi). Features at-risk graduation scoring, AI study plan generation via AWS Bedrock, multi-turn AI readiness checks, conversational course assistance, and teacher tools for course descriptions.
RAG-MT & QA — Finnish Culture Knowledge Base
Domain-specific retrieval-augmented generation for English-to-Finnish machine translation and question answering over a Finnish culture knowledge base. Evaluates baseline MT, naive RAG, and filtered context selection with BLEU, COMET, and entity novelty metrics. Achieves 95% QA accuracy.
Fake News Detection (Bangla)
A machine learning system for detecting fake news in Bangla language. Implements imbalance handling techniques and model stacking to improve classification accuracy on low-resource datasets.
Technical Skills
Technologies and tools I work with across the full stack