Karthik Ganesan
Research Scientist at Ambassador Software
San Francisco, California, United States
Summary
👤
Senior🎓
Top SchoolKarthik Ganesan is a Research Scientist based in San Francisco with 8 years of experience building and shipping applied ML systems across speech, multimodal, and LLM domains. Currently at Snowflake, he focuses on RL post-training, environments, and VoiceAI, and previously led LLM research for autonomous medical coding and founded AI startups and products. A CMU ML grad and former Alexa Applied Scientist, he has published on adapters for SSL speech models and contributed to the popular espnet end-to-end speech toolkit, where his commits touched core transducer blocks. He has a track record of taking research into production—deploying multilingual voicebots, improving SLU and ASR rescoring, and optimizing parameter-efficient serving with LoRA and self-refine techniques. Karthik combines academic rigor from CMU with hands-on startup pragmatism, and he often bridges long-context LLM research with production constraints. Outside work he invests in early-stage CX GenAI and mentors founders, reflecting a mix of technical depth and entrepreneurial curiosity.
8 years of coding experience
6 years of employment as a software developer
Bachelor of Engineering - BE Computer Science, Bachelor of Engineering - BE Computer Science at Ramaiah Institute Of TechnologyMasters in Intelligent Information Systems Machine Learning and Natural Language Processing, Masters in Intelligent Information Systems Machine Learning and Natural Language Processing at Carnegie Mellon University
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (5)
pytorch10
speech-recognition10
deeplearning-ai9
deep-learning9
machine-translation7
Programming languages (6)
Github contributions (5)
espnet/espnet
Nov 2021 - Jul 2022
End-to-End Speech Processing Toolkit
Role in this project:
ML Engineer
ML EngineerContributions:26 reviews, 45 commits, 13 PRs in 8 months
Contributions summary:Karthik's commits primarily involve modifications to the `espnet/espnet` repository, which is an end-to-end speech processing toolkit. The changes include merging branches and modifying the `espnet/nets/pytorch_backend/transducer/blocks.py` file, indicating a focus on adjusting network architectures or functionalities. The edits suggest contributions related to model training or modifications, potentially related to the core speech processing tasks the repository addresses.
speech-recognitionspeech-separationchainerspoken-language-understandingspeech-processing
skit-ai/Multimodal-Slu
Mar 2021 - May 2021
This repo builds an top of self-supervised speech embeddings using S3PL tool-kit and Text based transformers from Huggingface to explore multi-modal SLU
Contributions:20 commits, 1 PR, 13 pushes in 1 month
pytorchnlpmulti-modaltransformersmodal
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.