I'm a Software Engineer with 1+ year of experience in AI and Machine Learning, focusing on building scalable intelligent systems and conducting research on data-driven methodologies and model performance.
I work at the intersection of building and asking questions — shipping production software while contributing to applied machine-learning research.
I'm Abdul Jalil Tamjid, an engineer working in AI and machine learning, based in Dhaka, Bangladesh. Over the past year, I have worked on building practical systems and exploring applied research in areas like computer vision, deep learning, and intelligent systems.
My work focuses on understanding how models perform in real-world settings—from data preprocessing and model design to deployment and evaluation. I am particularly interested in building systems that are not only accurate, but also reliable, interpretable, and aligned with real-world impact.
Beyond my core work, I explore new ideas through hands-on experimentation, write about machine learning and systems, and enjoy helping others learn and grow in the field.
Designing and shipping reliable backend systems, ML pipelines, and developer tools that real teams depend on every day.
Investigating practical problems in NLP, distributed ML, and model evaluation — and writing them up for venues that actually publish code.
Teaching what I learn through essays, talks, code reviews, and the occasional 1:1 with a junior engineer trying to figure out their next move.
A short tour through the labs and engineering teams that shaped how I think about software and research today.
Worked on retrieval-augmented generation and evaluation methods for long-context models. Shipped a grounded-QA dataset adopted by three university courses; first author on three peer-reviewed papers.
Built the real-time fraud-detection backend serving 11M+ users. Migrated a 14-service legacy stack to four well-typed services. Owned the on-call rotation and the post-incident writeups nobody else wanted.
Graduated top 5% of class. Undergraduate thesis on transformer compression for low-resource languages, later extended into a workshop paper.
Postgraduate thesis on transformer compression for low-resource languages, later extended into a workshop paper.
My research sits at the seam of systems and machine learning: how we build, serve, and evaluate models at scale.
Designing and developing intelligent agentic systems using LangChain, LangGraph, and MCP, focusing on multi-step reasoning, tool integration, and autonomous decision-making workflows.
Building NLP-powered applications with emphasis on text generation, retrieval-augmented systems, and efficient processing of real-world textual data across diverse domains.
Developing and maintaining robust pipelines for text data generation, preprocessing, and deployment, ensuring scalability, reliability, and continuous integration of machine learning systems.
Open-source tools, internal platforms, and weekend experiments — a few favourites from the last few years.
A small thing built for an ecology lab in Tanzania that turned into a thing. Now deployed across 12 conservation projects.
Essays on systems, machine learning, and the strange job of being paid to think for a living.
Accuracy can create a false sense of success—especially in imbalanced datasets. This article explores why relying solely on accuracy can fail real-world models and how metrics like precision, recall, and AUC provide deeper insights.
Read articleEver wondered what happens behind the scenes when you book a train ticket, scroll through an app, or log into a website? This article breaks down APIs, monolithic architecture, and how modern applications communicate — with real-world examples.
Read articleWhy my favourite engineering output of last year was a 280-line script nobody asked for.
Read articleA modest reading system built around index cards, anchor links, and a single ugly script.
Read articleWhat it actually took to break into ML research from a backend engineering background.
Read articleTwelve books, three papers, one cookbook. None of them are about software, and that is the point.
Read articleHave a project, a research collaboration, or just want to chat about systems and ML? My inbox is open.
