•
Data and dashboards are not a strategy
Talks about Data Science, Data Engineering and Regression Modeling
RL
Richard Li
Post Train, Eval and Platform at TikTok
San Francisco Bay Area
Overview
Richard Li is a Generative AI expert currently serving as an Engineer Lead at TikTok, leading the GenAI Post Train, Eval, and Platform team. With a background in product development and data analysis, he has held senior management roles at Adobe and ASML, specializing in machine learning and software engineering.
Highlights of Richard's career include his leadership in developing AI-powered products at TikTok and Adobe, as well as his expertise in implementing machine learning solutions at ASML.
Work Experience
Engineer Lead: GenAI Post Train, Eval and Platform
2024 - Current
I lead the Generative AI Post-Training and Platform team, responsible for developing and fine-tuning various generative AI models and building an eval ecosystem (including text-to-image, text-to-video, Ads LLM, co-pilot, etc.) to support TikTok’s monetization solutions.
Data Platform Leader: Adobe Firefly
2022 - 2024
The effectiveness of LLM is limited by the quality of the data. I supervise teams responsible for curating various data modalities (text/images/videos/3D/audio, and behavioral data etc.). My focus is on establishing a resilient data framework that can efficiently handle and process petabytes of data pathways to enable data monetization and commercialization.
Senior Manager: Adobe Sensei
2019 - 2022
Deliver a machine learning solution for real time customer data platform.
Adobe Fund for Design,a $10m fund that supports emerging companies to build products and services
Senior Manager: Machine Learning
2017 - 2019
Building mass scale anomaly detection for on-premise/on cloud solution.
Engineering Manager: Machine Learning
2016 - 2017
Setting up machine learning framework across platforms for real-time image processing.
Engineering Manager, Software
2013 - 2016
Building machine learning solutions for defect inspection.
Senior Software Engineer
2010 - 2013
Building on promise solutions for defect detection on large scale SEM images.