Hi, I'm Safayet Hossain Masum. A Software Engineer with 4 years at Google and Amazon, specializing in high-throughput distributed systems, autonomous AI agents, and engineering workflow automation.
Scale, Reliability & Legal Compliance
Legally sensitive emails delivered
Yearly earnings impact at Amazon
Petabytes saved in consent storage
Infra cost reduced by redesign
I am a Software Engineer with 4 years of experience across Google and Amazon, specializing in building high-availability backend infrastructure, distributed scaling, and building production AI solutions.
Throughout my tenure in Big Tech, I have consistently navigated high ambiguity to deliver massive scale. At Amazon, I led technical execution for country-wide product releases, re-architected services to slash infrastructure costs by 75%, and contributed to scaling customer volume by 20x.
At Google, I architect mission-critical infrastructure that handles legal consents, automated translation review workflows, and coordinated high-volume user campaigns across the globe.
I am deeply passionate about applied Artificial Intelligence & Agentic Automation. In my daily engineering work, I actively build custom AI agents, orchestrate complex LLM workflows, and automate repetitive tasks—from code intelligence and regression analysis to operational tooling. I continuously master new AI frameworks to drive efficiency and build intelligent, resilient systems.
B.Sc. in Computer Science & Engineering
Problem Solving & Algorithmic Excellence
Active competitor on Codeforces, CodeChef, Toph, and StopStalk. Former Programming Coach at Green University of Bangladesh and National Level Contest Judge.
A summary of my education, academic research publications, contest records, and community talks.
Transitioned to Google Munich as Software Engineer after extensive tenure at Amazon Madrid as Software Development Engineer. Architected petabyte-scale storage, 150B+ email campaign infrastructure, and automated legal review pipelines.
Graduated with a B.Sc. in Computer Science & Engineering from the prestigious Islamic University of Technology (IUT). Featured by the IUT community and academic press as a celebrated graduate representing Bangladesh in Tier-1 global tech.
Veteran competitor across Codeforces, UVa Online Judge, Toph, AtCoder, and LeetCode. Contender in collegiate programming contests (including BRACU Techfest and ICPC regionals). Served as Programming Coach at Green University of Bangladesh and national contest judge.
Frequently invited guest speaker and mentor for tech sessions and university programming clubs (IUTCS, Uttara University CPC). Delivered workshops including "Building a Professional Portfolio with Git & GitHub" and career guidance webinars on cracking Big Tech interviews.
Author / creator of the published "HybridDataset for Bengali Image Captioning" published on Mendeley Data. Research focused on deep neural networks, CNN-LSTM models, and assistive computer vision technologies designed to aid visually impaired individuals.
Active builder on GitHub (Safayet-masum), developing developer tooling, autonomous AI workflows, automated test pipelines, and distributed utilities. Passionate about empowering engineering communities through open architectures.
A timeline of my software engineering roles across Google, Amazon, and university coaching.
Software Engineer
Jul 2025 – Present
Munich, Germany 🇩🇪
Software Development Engineer
Oct 2022 – Jun 2025
Madrid, Spain 🇪🇸
Programming Coach
Jun 2022 – Sep 2022
Dhaka, Bangladesh 🇧🇩
Mentored university students on Data Structures, Algorithms, Graph Theory, and Competitive Programming utilizing C++.
Tools, languages, and frameworks I use in my day-to-day engineering and side projects.
A selection of academic research, systems software, and development tools I've built.
In-browser AI execution engine running Llama-3-1B and local LLMs directly on consumer GPUs via WebGPU compute shaders. 100% private with zero backend API requests and persistent local caching.
Designed and trained a novel ML model to facilitate spatial navigation for visually impaired individuals through real-time scene captioning. Later built a Bengali captioning model achieving state-of-the-art results.
Constructed a comprehensive automated code judge for university programming contests and lab evaluations. Includes sandboxed compilation, offline problem solving mode, performance metrics storage, and automated plagiarism detection.
Deep-dives into distributed architectures, computer vision research, and problem solving.
Architectural insights, queuing mechanisms, and legal compliance safeguards when executing hyper-scale distributed campaigns at Google and Amazon.
How we leveraged deep learning architectures, attention mechanisms, and dataset curation to create state-of-the-art assistive captioning models.
How years of algorithmic contests on Codeforces, CodeChef, and university coaching shaped my approach to software engineering at Google and Amazon.
I'm always happy to talk about engineering challenges, backend architecture, AI workflows, or new opportunities.
Direct Email
safayet@safayetmasum.comCurrent Base
Munich, Germany 🇩🇪
Status
Open to AI initiatives
Leave your details and message below; I'll get back to you promptly.
Thank you for reaching out, Safayet will review your inquiry.
Ask me anything about Safayet
