Dexter J
Member Of Technical Staff at Microsoft AI
New York, New York, United States
Summary
🤩
Rockstar🎓
Top SchoolDexter J is a research-focused software engineer with a decade of experience building and scaling large language model systems and safety infrastructure, currently working on AI safety at Microsoft AI. Previously a core member of Meta’s LLaMA and FAIR teams, he led post-training reasoning, RL infrastructure, and multi-modal conversational agent work that contributed to models like BlenderBot and publications at top-tier conferences. He bridges deep research and production engineering, having shipped dataset integrations and QA tasks to the widely used ParlAI framework and addressed transformer-specific metric and reliability issues. Dexter’s background spans NLP, reinforcement learning, and systems scaling, complemented by early quantitative and engineering roles in finance and embedded systems. Based in New York, he combines rigorous academic training from Télécom Paris and Sorbonne with hands-on expertise in Python and C++ to move frontier LLM research toward safe, deployable outcomes. An underappreciated strength is his track record of improving community-facing tooling—making cutting-edge research reproducible and robust in open-source ecosystems.
10 years of coding experience
8 years of employment as a software developer
Master's degree Data Science, Master's degree Data Science at Sorbonne UniversitéMaster's degree Data Science and Human Computer Interaction, Master's degree Data Science and Human Computer Interaction at Télécom Paris
English, French
Github Skills (9)
data-set10
transformer-models10
machine-learning10
nlp10
python10
data-integration10
conversational-ai10
datasets10
pytorch9
Programming languages (3)
Github contributions (5)
facebookresearch/ParlAI
May 2019 - Nov 2022
A framework for training and evaluating AI models on a variety of openly available dialogue datasets.
Role in this project:
ML Engineer
ML EngineerContributions:19 reviews, 72 commits, 104 PRs in 3 years 6 months
Contributions summary:Dexter contributed significantly to the `parlai` repository, focusing on the integration and improvement of conversational question-answering (CoQA) and question-answering in context (QuAC) datasets. Their work involved creating and modifying task-specific agents, build files, and task lists to incorporate these datasets. Additionally, they added and modified metrics, and addressed potential issues with transformer models.
nlpdeep-learningdatasetmachine-learningtraining
dexterju27/dexterju27.github.io
Feb 2019 - Feb 2024
Contributions:18 pushes, 1 branch in 5 years
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.