SP
AI Engineer · Full-Stack Software Engineer

Field note // 62.24°N 25.75°E · JYU — MSc AI · Systems trace — retrieval → reasoning → shipping

MD. SALMAN HOSSANPROTTOY

AI Engineer/Full-Stack Software Engineer

Software Engineer with 4+ years building production full-stack, backend, and AI-enabled systems end-to-end — including a production RAG platform with document ingestion, multi-LLM retrieval, vector search, caching, and Docker-based deployment. MSc Artificial Intelligence at University of Jyväskylä.

I ship from backend architecture through to the interface. If the retrieval fails, the UI shouldn't lie — and if the migration is risky, it ships with zero downtime.

Jyväskylä, Finland · +358 44 955 0788Open to AI & Full-Stack opportunities
Let's talkResumeExplore the work
Md. Salman Hossan Prottoy

Builder profile

Md. Salman Hossan Prottoy

Jyväskylä, Finland · AI Engineer · human in the loop

Live retrieval trace ● interactiveembedding…
↩
queryembeddingvector searchLLM
Vector field · 9 docs · cosine2 retrieved
DOC·014
DOC·089
DOC·112
DOC·203
DOC·041
DOC·077
DOC·156
DOC·198
DOC·063

DOC·014

Opin.fi / CS curricula

DOC·089

Brand Cloud RAG corpus

Generated · grounded in retrieved docs

Sources:DOC·014DOC·089

Grounded in a local index of this portfolio — ask your own question.

About — Field ledger

Systems that survive contact with users.

How I work: retrieval that grounds, infrastructure that holds, and interfaces that don't lie.

Software Engineer with 4+ years building production full-stack, backend, and AI-enabled systems end-to-end — including a production RAG platform with document ingestion, multi-LLM retrieval, vector search, caching, and Docker-based deployment. Pursuing an MSc in Artificial Intelligence at the University of Jyväskylä, Finland.

Skilled across Python, TypeScript, Go, Java, React/Next.js, LLM integration, vector search, and cloud infrastructure, with a track record of shipping reliable, scalable systems from backend architecture through to client-facing features. Core stack: Python, TypeScript, Go, Java, React, Next.js, Node.js, PostgreSQL, Redis, AWS (Lambda, RDS, SQS), Docker, RAG, LLMs, LangChain, Vector Search, TensorFlow.

Build for the happy path. Design for everything that happens after it.

Selected signals // ledger

93.1%

F1 on published Bangla fake-news study

0

downtime — CMS to PostgreSQL on RDS

40%

of travel platform API endpoints shipped

4+

years production experience

Source: production telemetry · Brand Cloud / W3 / JYU
Experience — Service record

Where I have shipped.

Production systems, zero-downtime migrations, and product work · Japan ↔ Bangladesh · 2021–2026

Software engineering · full stack · infrastructure · applied AI

Software Engineer

Brand Cloud Inc. · Tokyo, Japan (Remote)
Sep 2023 — Jun 2026
  • Built a production RAG application with document ingestion, multi-LLM integration, retrieval, vector search, caching, and Docker-based deployment.
  • Developed and maintained full-stack applications using TypeScript, React, and Next.js across 3+ major projects, with production deployment on Vercel and measurable increase in user satisfaction scores.
  • Migrated CMS data storage to PostgreSQL on Amazon RDS with zero downtime, improving reliability and simplifying production data management.
  • Selected tickets each sprint, assigned tasks to a team of 5 engineers, and improved sprint delivery times; conducted 100+ code and screen reviews, maintaining high code quality and reducing bugs.
  • Automated GitHub-based CI/CD pipelines and improved software delivery workflows using DevSecOps practices.
  • Mentored 2–3 junior engineers and contributed to coding standards, architecture discussions, and code reviews.

Junior Software Engineer

W3 Engineers Limited · Dhaka, Bangladesh
May 2022 — Aug 2023
  • Designed and developed scalable, robust REST and GraphQL APIs using Go and Node.js, contributing to ~40% of API endpoints across a multi-tenant travel platform covering 30+ domains.
  • Contributed to migrating backend services from Django to Go, improving API performance and reducing response latency.
  • Developed, maintained, and optimized multiple cron jobs for data processing and updates, using parallel processing techniques to boost efficiency.
  • Built data pipelines integrating multiple external providers into a unified schema, using Amazon SQS for asynchronous workloads on AWS.
  • Developed migration scripts and integrated Elastic File System (EFS) for data storage, facilitating seamless database transitions and enhancing file access speed.

Software Engineer Intern

W3 Engineers Limited · Dhaka, Bangladesh
Nov 2021 — May 2022
  • Designed reusable, scalable Vue.js components for an enterprise network-monitoring application, powering ~35% of the frontend.
  • Improved application performance and reliability through regular performance testing and bottleneck identification.
  • Introduced Vue.js, Flask, and Beego; built REST APIs with Beego, Swagger UI, and PostgreSQL, web scraping with Flask and MySQL, and data visualization with Chart.js, Vuex, and InfluxDB.

Undergraduate Teaching Assistant

American International University-Bangladesh · Dhaka, Bangladesh
May 2021 — Aug 2021
  • Provided academic support in the Web Technologies course for ~80 students; collaborated on lesson planning, instructional materials, and student progress reporting.
  • Assisted with grading, lab guidance, and individual student support.
Education — Research ledger

Always still learning.

The academic foundation behind intelligent systems — thesis in multi-objective optimization.

Aug 2025 — Jun 2027 (Expected)

MSc Artificial Intelligence

University of Jyväskylä · Jyväskylä, Finland

Agent Technologies for Developers · Deep-Learning for Cognitive Computing · SOA and Cloud Computing · Collective Intelligence and Agent Technology · Semantic Technologies

Jan 2018 — Sep 2021

BSc Computer Science & Engineering

American International University-Bangladesh · Dhaka, Bangladesh

Data Structure · Algorithms · OOP · Introduction to Database · Artificial Intelligence · Data Warehousing and Data Mining · Web Technologies

Recognition — Peer review

Awards & Publications.

Research that shipped beyond the lab — peer-reviewed and awarded.

Mar 2023

Vice Chancellor's Award, AIUB

Awarded for the published Bangla fake-news thesis project.

ICFIR 2021, LNNS 437, Springer (2022)

Approaches for Improving the Performance of Fake News Detection in Bangla: Imbalance Handling and Model Stacking

Hossain, M.M., Awosaf, Z., Prottoy, M.S.H., et al.

50K-instance BanFakeNews dataset (97% majority skew) · 93.1% F1 with SMOTE vs 67.6% baseline; 79.1% with stacked generalization

Read publication
English:Professional Working ProficiencyFinnish:Elementary Proficiency
Toolkit — Instrumentation

How I approach the work.

A practical mix of product thinking, systems craft, and curiosity — one pipeline, end to end.

SHIP

Production-minded

From zero-downtime PostgreSQL on RDS to GitHub CI/CD with DevSecOps, I care about what keeps software dependable after launch.

EXPLORE

Applied AI

RAG, multi-LLM retrieval, vector search, CrewAI multi-agent systems — building LLM experiences that survive real usage, not just demos.

CONNECT

Team multiplier

Sprint planning for 5 engineers, mentoring 2–3 juniors, and turning lessons into reusable components and clear standards.

Contact — Open channel

Have a hard problem?Let's make it useful.

I'm always interested in thoughtful product teams, ambitious engineering challenges, and applied AI that earns its place in the workflow.

Jyväskylä, Finland · salman.prottoy@gmail.com · +358 44 955 0788

salman.prottoy@gmail.com

Usually replies within a day. Thesis-friendly hours.