Sabrina Mielke
Machine Learning Engineer, Large Language Models at Genentech
New York, New York, United States
Summary
🤩
Rockstar🎓
Top SchoolSabrina Mielke is a Machine Learning Engineer specializing in large language models with 12 years of experience bridging academic research and industry applications. She holds a PhD from Johns Hopkins and has applied LLMs to biotech drug discovery at Genentech after roles at AlphaSense, Cohere, and internships at Google AI, FAIR, and Hugging Face. Her work spans research, production engineering, and teaching—she designed and taught an AI course at JHU and contributed to the ParlAI dialogue framework, improving dataset pipelines for Squad2, TriviaQA, and Wizard of Wikipedia. Comfortable in both research and product contexts, she has published and collaborated on tokenization and conversational calibration research while shipping practical model tooling. Based in New York, she combines deep NLP expertise with hands‑on data and training pipeline work, often focusing on making models safer and better calibrated. An understated strength is her history of building reproducible dataset and preprocessing fixes that quietly enable more reliable downstream model evaluation.
12 years of coding experience
7 years of employment as a software developer
Johns Hopkins UniversityMaster of Science - MS Computer Science, Master of Science - MS Computer Science at Technische Universität Dresden
English, German, Latin, Chinese, Arabic, Japanese
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (8)
data-preprocessing10
machine-learning10
nlp10
python10
data-integration10
natural-language-processing10
datasets10
data-set10
Programming languages (14)
Github contributions (5)
facebookresearch/ParlAI
May 2020 - Apr 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:8 commits, 13 PRs, 12 pushes in 1 year 10 months
Contributions summary:Sabrina contributed to the ParlAI framework by implementing and modifying components related to various dialogue datasets and model training. They fixed data setup issues for the Squad2 dataset, integrated no-evidence and union functionalities for TriviaQA, and updated the TriviaQA build process. Additionally, the user made changes to incorporate dummy knowledge in the Wizard of Wikipedia project and updated checksums for the NarrativeQA dataset, indicating involvement in data preprocessing and task-specific configurations.
nlpdeep-learningdatasetmachine-learningtraining
sjmielke/pcfg-experiments
Feb 2017 - Jul 2017
Contributions:143 commits, 81 pushes, 1 branch in 4 months
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.