Jiatong Gao
Machine Learning Engineer at Affirm
Pittsburgh, Pennsylvania, United States
Summary
🤩
Rockstar🎓
Top SchoolJiatong Gao is a Machine Learning Engineer with nine years of experience specializing in speech processing and representation learning, now based in Pittsburgh and currently at Affirm. He has contributed substantial backend and full-stack code to major open-source toolkits like s3prl and ESPnet, focusing on speaker diarization, ASR recipe development, and vocoder integration. Jiatong’s work blends research and production engineering—he built diarization datasets and core data structures for real-world pipelines while also tuning model architectures and training scripts. A Carnegie Mellon MS in Information Security and a NYU CS BA underpin his cross-disciplinary approach to ML system design, where security-aware thinking informs reliable model deployment. His contributions to well-known projects in the speech community show both deep domain expertise and a knack for turning research components into reproducible tooling.
9 years of coding experience
Bachelor of Arts - BA, Computer Science, 3.8/4.0, Bachelor of Arts - BA, Computer Science, 3.8/4.0 at New York UniversityMaster of Science - MS, Information Security, 3.9/4.0, Master of Science - MS, Information Security, 3.9/4.0 at Carnegie Mellon University
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (15)
speaker-diarization10
pytorch10
machine-learning10
text-to-speech10
bash10
audio-processing10
speech-recognition10
speech-synthesis10
automatic-speech-recognition10
python10
asr10
data-processing10
datasets10
data-set10
realtime7
Programming languages (6)
Github contributions (5)
espnet/espnet
Sep 2019 - Jan 2023
End-to-End Speech Processing Toolkit
Role in this project:
Back-end Developer
Back-end DeveloperContributions:651 reviews, 922 commits, 466 PRs in 3 years 4 months
Contributions summary:Jiatong's contributions focused on enhancements to the egs2/chime7_task1 repository, including merging master branches and adding code to gen_task1_data.py. These changes involved code modifications related to audio processing scripts and the creation of speech recognition recipes. The user also set none for ctc weights within the espnet2/asr/espnet_model.py file, indicating involvement in model architecture adjustments, specifically for automatic speech recognition (ASR) tasks.
speech-recognitionspeech-separationchainerspoken-language-understandingspeech-processing
kan-bayashi/ParallelWaveGAN
Jan 2022 - Jan 2023
Unofficial Parallel WaveGAN (+ MelGAN & Multi-band MelGAN & HiFi-GAN & StyleMelGAN) with Pytorch
Role in this project:Full-stack Developer
Full-stack DeveloperContributions:14 commits, 9 PRs, 11 comments in 11 months
Contributions summary:Jiatong's contributions involve adding recipes and related scripts for different speech synthesis and vocoder tasks. This includes setting up data preparation and processing pipelines, as well as integration with vocoder training and decoding scripts. The user demonstrates an understanding of the project's structure by creating example configurations and integrating new datasets.
realtimebandneural-vocodermelganparallel-wavenet
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.