It's me

Hey! I am

Hey! I am

Parvez M Robin
Parvez M Robin

(He/Him)

(He/Him)

I make machines intelligent

I make machines intelligent
It's me

parvezmrobin[at]gmail[dot]com

As a polyglot software engineer with a strong background in software development and machine learning, my experience includes firsthand work with C++, Go, Python, and TypeScript for over three years.

Currently, I am a machine learning engineer at Siemens EDA AI – creating next generation AI platform for Electronic Design Automation (EDA).

Skills
I can exit Vim. In addition to that, I am skilled in
Languages
Python,
C++,
Go,
JavaScript (ES6+),
TypeScript,
C#,
PHP
Artificial Intelligence
Neural Networks,
Transformers,
Large Language Models,
Retrieval Augmented Generation,
Transfer Learning,
Ensemble Learning,
Unsupervised Learning,
Natural Language Processing
Web Technology
Node JS,
Vue JS,
React,
REST API,
Bootstrap,
CSS3,
Flask,
FastAPI
Database
Mongo DB,
MySQL,
PostgreSQL,
Redis
Tools & Platform
Docker,
Google Cloud Platform,
Amazon Web Service,
Microsoft Azure,
Linux
Experience
Siemens
Software Developer March 2024 — Present

Siemens

Software Developer

March 2024 — Present

  • For Analog FastSPICE – the fasted nanometer circuit verification platform:
    • Developed an element stamping algorithm that takes up to 4 times less memory in multithreaded runs.
    • Jointly developed a Matrix-Vector multiplication algorithm specific to symmetric matrices that is 40% faster than the general-purpose algorithm.
  • For Siemens EDA AI:
    • Developed a hybrid rerank RAG pipeline for searching relevant documents using both semantic and keyword-based search (ElasticSearch & Milvus).
    • Architected and developed an Auth service that facilitates multiple authentication backends (e.g., OAuth, LDAP) and multiple authorization schemes (RBAC, ABAC).
Metabob
Research Intern May 2022 — April 2023

Metabob

Research Intern

May 2022 — April 2023

  • Worked on a research project aiming to automatically explain software bugs in natural language using cutting-edge deep learning techniques. A recent developer survey shows that our developed tool can provide 43% more accurate and 40% concise explanations of bugs (see Publications).
  • Developed a web application backed by state-of-the-art language model to explain defective source code using Python, Flask, PyTorch, and Vue JS (see Demo).
NerdDevs
Software Engineer March 2019 — July 2021

NerdDevs

Software Engineer

March 2019 — July 2021

  • In Daency, an online platform to learn dancing interactively:
    • Led the architecture and development where used TypeScript, Express JS, Vue JS, Mongo DB, and Docker as the tech stack.
    • Integrated and maintained WebRTC, AWS (S3 + Cloudflare), GCP (Authentication), Push Notification, Stripe, Task Queue (redis) and various other services.
    • Collaborated with the stakeholders to engineer requirements and plan sprints in the agile process with a globally distributed team, while reducing requirement changes during sprints from ≈ 30% to < 5%.
  • In Bikribatta inventory solution:
    • Architected accounts, reports, and employee management components.
    • Written a stock management microservice using Golang that reduces the latency 10X.
    • Built a desktop app for the Point of Sales component using Electron, which also works in offline mode with an eventual consistency scheme.
  • Migrated the subscription backend of Wonster Words iOS app store game from node.js to ASP.Net using Microsoft Azure FaaS, Cosmos DB, and Queue Storage.
  • Contributed to several in-house React Native based mobile applications.
  • Hosted two in-house workshops on ‘Modern Web Development with Vue JS’ and ‘Sustainable Architecture with TypeScript’.
Festive.Rocks
Co-founder and Tech Lead January 2023 — Present

Festive.Rocks

Co-founder and Tech Lead

January 2023 — Present

  • Created a low latency, high-performance website to run on slow internet regions for the target market.
  • Seamlessly hosted 229 events with great customer satisfaction.
  • Led a team of four developers to complete a fully featured web application within three months while maintaining very high code quality and more than 90% test coverage.
Publications

Explaining Software Bugs Leveraging Code Structures in Neural Machine Translation

Parvez Mahbub, Ohiduzzaman Shuvo and M. Masudur Rahman, In Proceeding of The 45th IEEE/ACM International Conference on Software Engineering (ICSE), 2023, pp. 640–652, Melbourne, Australia, May 2023
Software Bug,Bug Explanation,Software Engineering,Software Maintenance,Natural Language Processing,Deep Learning,Transformers
Google Scholarpre-printdemo

Predicting Line-Level Defects by Capturing Code Contexts with Hierarchical Transformers

Parvez Mahbub and M. Masudur Rahman, In Proceeding of The 31st IEEE International Conference on Software Analysis, Evolution and Reengineering (SANER), 2024, pp. 12, Rovaniemi, Finland, March 2024 (to appear)
Software Quality Assurance,Line Level Defect Prediction,Code Context,Code Structure,Software Engineering,Software Maintenance,Deep Learning,Transformers
Google Scholarpre-print

Authorship Identification of Source Code Segments Written by Multiple Authors Using Stacking Ensemble Method

Parvez Mahbub, Naz Zarreen Oishie and S. M. Rafizul Haque, In Proceeding of the 22nd International Conference on Computer and Information Technology (ICCIT), 2019, pp. 1-6
Source Code Authorship Identification,Deep Neural Network,Random Forest,Support Vector Machine,Stacking Ensemble
Google Scholarpre-print

Defectors: A Large, Diverse Python Dataset for Defect Prediction

Parvez Mahbub, Ohiduzzaman Shuvo and M. Masudur Rahman, In Proceeding of The 20th Mining Software Repositories (MSR), 2023, pp. 393-397, Melbourne, Australia, May 2023
Defect Prediction,Just-in-Time,Dataset,Software Engineering,Software Maintenance
Google Scholarpre-print

Bugsplainer: Leveraging Code Structures to Explain Software Bugs with Neural Machine Translation

Parvez Mahbub, M. Masudur Rahman, Ohiduzzaman Shuvo and Avinash Gopal, In Proceeding of The 39th IEEE International Conference on Software Maintenance and Evolution (ICSME 2023), 2023, pp. 6, Bogotá, Colombia, October 2023
Software Bug,Bug Explanation,Software Engineering,Software Maintenance,Natural Language Processing,Deep Learning,Neural Text Generation
Google Scholarpre-printdemo

Improved Detection and Diagnosis of Faults in Deep Neural Networks Using Hierarchical and Explainable Classification

Sigma Jahan, Mehil B. Shah, Parvez Mahbub and M. Masudur Rahman, In Proceeding of The 47th International Conference on Software Engineering (ICSE), 2025, pp. 13, Ottowa, Ontario, Canada, April 2025 (To appear)
Software Engineering,Deep Neural Networks,Dynamic Analysis,Model Fault,Static Analysis,Training Fault
pre-print

Recommending Code Reviews Leveraging Code Changes with Structured Information Retrieval

Ohiduzzaman Shuvo, Parvez Mahbub and M. Masudur Rahman, In Proceeding of The 39th International Conference on Software Maintenance and Evolution (ICSME), 2023, pp. 194-206, Bogotá, Colombia, October 2023
Software Engineering,Code Reviews,Code Changes,Information Retrieval,Software Maintenance
Google Scholarpre-print
Activities
  • [May 2022 — Apr 2024]As the Vice President External Advocacy in Dalhousie Association of Graduate students (DAGS), maintained relations with several external organizations and committees including Department of Advanced Education, Nova Scotia
  • [May 2022 — Apr 2023]As the Graduate Student Representative in Dalhousie Student Union (DSU) Council, ensured the decisions and policies passed in the DSU are in favor of the graduate students
  • [May 2022 — Apr 2023]As the Graduate Student Representative in Faculty of Graduate Studies (FGS) Council of Dalhousie University, I ensured the students’ interest in faculty policy
  • [2017 — 2018]As the Head of Promotion of Podium, arranged a dozen sessions Khulna University — an organization for improving public speaking skills
  • [2018]Served as the Head of the Souvenir Committee in Khulna University Central Rag, 2018
  • [2018]Managed official website of CSE discipline, Khulna University, and corresponding computer club — CLUSTER as the IT secretary of CLUSTER.
Talks & Presentations
Education
school image
Master of Science in Computer Science
Dalhousie University

September 2021 – August 2023

Thesis: Comprehending Software Bugs Leveraging Code Structure in Neural Language Modelling
Supervisor: Masud Rahman

Bachelor of Science in Computer Science and Engineering
Khulna University

January 2015 – February 2019

Thesis: Authorship Identification of Source Code Segments Written by Multiple Authors Using Stacking Ensemble Method
Supervisor: S.M. Rafizul Haque

school image
Personal Projects

Jhijhi

A react app backed by Express JS and Mongo DB to facilitate cricket scoring with heavy customization support. It is hosted with Netlify (frontend), Amazon EC2 (backend), MongoDB Atlas (database), and Circle CI (deployment).
NodeJS Express React JavaScript MongoDB
Jhijhi

Visually Explaining Bug Localization

Developed a visualization system for a state-of-the-art bug localization tool (LR+WE by Ye et al.). It visually explains the localization process gradually at document-to-document, document-to-word and finally word-to-word level.
Python VueJS D3JS TypeScript Flask scikit-learn Word Embedding NLTK
Visually Explaining Bug Localization

Bug Report Deduplication Using Siamese Cross-Encoder Network

Replicated a state-of-the-art duplicate bug report detection technique based on a siamese architecture. We showed that even a simpler version of the original work can achieve comparable performance.
MongoDB Docker TensorFlow scikit-learn
Bug Report Deduplication Using Siamese Cross-Encoder Network

Official Website of CSE Discipline, Khulna University

Led team to renovate the official website of the Compute Science and Engineering Discipline, Khulna University, Bangladesh.
PHP Bootstrap MySQL
Official Website of CSE Discipline, Khulna University
Achievements
  • [Sep 2021 — Aug 2023] Dalhousie University Scholarship: Awarded by the Faculty of Graduate Studies and the Faculty of Computer Science for academic results and research excellence
  • [May 2022 — Apr 2023] Mitacs Accelerate Scholarship: Awarded by Mitacs to research in explaining software bugs using artificial intelligence
  • [Apr 2019] Placed among the top 15 teams in Datathon Bangladesh, 2019 by Robi Axiata Group
  • [Oct 2017] Champion in the programming contest in CSE Fest 2017, Khulna University, Bangladesh
  • [May 2017] 2nd runner-up in DIU App Contest, 2017 arranged by Daffodil International University, Dhaka, Bangladesh
  • [May 2015] Runner-up in the programming contest in CSE Fest 2015, Khulna University, Bangladesh
  • [Jan 2015 — Dec 2016, 2018] Khulna University Scholarship: Awarded by the Computer Science and Engineering Discipline and Khulna University scholarship committee based on academic excellence
Certifications
Testimonials
Parvez M Robin is a knowledgeable and insightful person who likes to share his thoughts and ideas on various topics related to software engineering, machine learning, and natural language processing. He writes blog posts on his personal website where he explains complex concepts in simple and humorous ways. He is also a learner and teacher who constantly updates his skills and knowledge by taking courses, reading books, and attending workshops. He also helps others learn by hosting workshops, taking classes, and providing mentorship.

— Bing Chat bing
For serious testimonials check out my LinkedIn profile I Wonder!
Want to know more?

Why don’t you send an email to parvezmrobin[at]gmail[dot]com and let’s talk. I Wonder!

This website is built with Vue 3, Bootstrap 5, TypeScript and Vite. It is hosted on Netlify. The source code is available on GitHub.