Md Awsaf Alam

AI Product Engineer · R&D Leadawsafalam@gmail.comawsafalam.com

linkedin.com/in/awsafalamgithub.com/awsafalam

Case studiesWriting

Professional Summary

AI product engineer and R&D lead with founding-engineer experience. I build agent-powered workflows and the systems needed to evaluate and operate them. At Sourcy, my work spans supplier conversations, internal product interfaces, and translating operational feedback into product and engineering decisions.

Experience

Sourcy Global Ltd.

– Present

R&D Lead

– Present
  • Rebuilt the supplier-conversation agent around explicit goals, conversation history, and human escalation, supported by end-to-end tracing and regression testing.
  • Integrated supplier intelligence into the existing sourcing interface, linking answers to supplier records and sharing a testable workflow with operations the same afternoon.
  • Audited production escalations across 55 sourcing requests, identifying missing briefs and category mismatches as upstream blockers and proposing intake validation and request-level triage.
  • Worked with operations and product teams to clarify image-search intent, pending-versus-failed states, and filtering trade-offs during a new sourcing-flow rollout.

Tech Lead

–
  • Delivered a social media insights platform for TikTok and Instagram, combining AI agents with continuous data pipelines.
  • Cut customer wait times for sourcing results by 90% through intelligent search and proactive data scraping.
  • Automated 70% of back-office sourcing processes and raised sourcing recommendation accuracy to 80%.
  • Scaled infrastructure with serverless functions to support a 30% month-over-month increase in transaction volume.
  • Reduced platform downtime by 25% after introducing monitoring.

Full Stack Engineer

–
  • Designed the backend for Sourcy.ai as modular microservices, with CI/CD pipelines and automated tests.
  • Built generative AI data-validation pipelines and a web-scraping API for data ingestion and enrichment, reducing manual work.
  • Cut database query response time by 40% through normalization, indexing and caching.

Senior Software Engineer [Part-Time], Boomershub LLC

–
  • Moved the frontend from a legacy React application to Next.js, improving SEO and web performance.

Founding Software Engineer, Sohopathi Pte Ltd

–
  • Designed the backend architecture for Sohopathi’s web and mobile applications, ensuring scalability and reliability.
  • Defined API specifications and led backend development, facilitating seamless integration across platforms.

Software Engineer, Mainframe Labs Ltd.

–
  • Built and maintained web applications in a full-stack JavaScript role.

Junior Software Engineer, Headless Technologies

–
  • Worked on backend services, dashboards and core chatbot architecture.

Education

MSc. Computer Science & Engineering, Bangladesh University of Engineering & Technology

2026

Specialization: Advanced Dependable and Fault-Tolerant Computer Systems, Network Security, Distributed Computing Systems, Software Quality Assurance

BSc. Computer Science & Engineering, Bangladesh University of Engineering & Technology

Publications & Research

  1. ToxiShield: Promoting Inclusive Developer Communication through Real-Time Toxicity Filtering
    Md Awsaf Alam Anindya, Showvik Biswas, Anindya Iqbal, Jaydeb Sarker, Amiangshu Bosu
    Proceedings of the ACM on Software Engineering (PACMSE) Vol. 3, FSE 2026, Main Research Track · DOI · arXiv
  2. Real-Time Toxicity Filtering for Open-Source Code Reviews
    Md Awsaf Alam Anindya, Showvik Biswas, Anindya Iqbal, Jaydeb Sarker, Amiangshu Bosu
    FSE 2026 Companion (2 pages) · DOI · arXiv
  3. Promoting Inclusive Developer Communication through Real-Time Toxicity Filtering
    M.Sc. Engg. thesis (ToxiShield), Department of CSE, BUET, 2026

Earlier research: automated code repair from code reviews (NLP) · COVID-19 contact-tracing app reviews · distributed databases in cloud-native environments

Full details: awsafalam.com/research/toxishield

Projects

GitMiner

githubMiner
  • A web crawler for detecting and gathering code in different repositories, and stores them locally
  • Tools Used: Python, Scrapy

Taint Analysis for Automated Code Repair

openrefactory.com
  • Analyzed Java libraries to build a dataset for training automated code correction models.
  • Tools Used: Java, SQL, Python

Web Automation Scraper

web-automation
  • Automated Pipeline to scrape data from various healthcare facility websites in the US, and a searchable interface for users to find those properties
  • Tools Used: Python, Nodejs, Sequelize, Docker

Additional Experience And Awards

  • Winner - MIT Solve, Tiger Challenge, 2019
  • Champion - GDG HackSprint, 2018
  • Champion - Tech For Peace Hackathon, 2016

Technologies

Languages: TypeScript, JavaScript, Python, SQL

Frameworks: Node.js, Express, React, Next.js

Data: Postgres, MongoDB

Infrastructure & AI: Docker, AWS, Cloudflare Workers, Generative AI, AI agents, LLM evaluation

Practices: CI/CD, microservices, object-oriented and functional programming, Agile