Pegah Ghahremani
AI Research Scientist at Meta
Seattle, Washington, United States
Summary
👤
Senior🎓
Top SchoolPegah Ghahremani is an AI research scientist with 13 years of experience building and deploying machine learning systems across industry leaders such as Amazon, Meta, AssemblyAI and Microsoft, and holds a PhD in Computer Engineering from Johns Hopkins. She specializes in speech and language technologies—spanning NLP, computer vision, recommender systems and anomaly detection—with hands-on expertise in Python, deep learning, and production-grade model engineering. At AssemblyAI she led multilingual SpeechLLM efforts and competitive benchmarking that materially improved conversational intelligence metrics, and she has contributed upstream to the widely used Kaldi ASR project by improving pitch extraction and performance. Known for bridging rigorous research with scalable production impact, she repeatedly reduces missed-entity and accuracy gaps in real-world systems. Based in Seattle, she combines academic depth with a track record of shipping robust, high-performance models in fast-moving teams.
13 years of coding experience
6 years of employment as a software developer
Johns Hopkins University
Persian, English, Arabic
Stackoverflow
13reputation
92reached
0answers
1question
Github Skills (11)
speech-to-text10
kaldi10
speech-recognition10
c-language10
speech10
cprogramming-language10
shell9
machine-learning8
python8
cuda7
tensorflow6
Programming languages (2)
Github contributions (5)
kaldi-asr/kaldi
Sep 2015 - Apr 2019
kaldi-asr/kaldi is the official location of the Kaldi project.
Role in this project:
Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:32 PRs, 271 comments in 3 years 8 months
Contributions summary:Pegah primarily contributed to the Kaldi project by adding and modifying pitch extraction code, including testing and experimentation with different configurations and parameters. They added test cases for pitch extraction functions and benchmarked against existing datasets. Further contributions included script fixes for pitch extraction, particularly in the babel example, and modifications to support different sampling rates and parameter variations. They also worked on speed improvements within the pitch-extraction codebase.
cudakaldiasrspeech-to-textkaldi-asr
pegahgh/kaldi
Sep 2015 - Mar 2019
This is now the official location of the Kaldi project.
Contributions:226 commits, 13 PRs, 288 pushes in 3 years 6 months
kaldilocation
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.