Song Ding
@songding
Senior AI Engineer building scalable multimodal ML systems from research to production.
United StatesWhat I'm looking for
I’m a Senior AI Engineer with 15 years of experience architecting and deploying scalable machine learning systems, specializing in computer vision, speech processing, and multimodal architectures. I orchestrate end-to-end model lifecycles—from data ingestion and distributed training to high-availability inference pipelines—serving millions globally while improving both model latency and precision.
At ByteDance, I architected a multimodal AI framework that increased user engagement by 25% and reduced end-to-end inference latency by 35% while improving Word Error Rate (WER) by 18% using advanced quantization in speech recognition pipelines. Earlier at Yahoo, I improved transcription accuracy from 87% to 95%, enabled real-time edge processing with TensorRT optimization, reduced training cycles by 40% with proprietary synthetic data augmentation, and built scalable MLOps pipelines. I’m passionate about advancing research-to-production by integrating large-scale generative models and proprietary neural architectures to tackle complex challenges and maintain competitive advantage.
Experience
Work history, roles, and key accomplishments
Architected a multimodal AI framework integrating vision transformers with large language models, increasing user engagement by 25%. Reduced end-to-end inference latency by 35% and improved WER by 18% using advanced quantization, while boosting throughput by 50% via PyTorch + TensorRT and reducing infrastructure costs by 20%.
AI Engineer
Yahoo!
Dec 2014 - Jun 2021 (6 years 6 months)
Architected a multi-modal enterprise platform unifying computer vision with ASR pipelines, improving workflow efficiency by 20%. Engineered transformer-based acoustic models to raise transcription accuracy from 87% to 95% and enabled real-time edge processing by reducing latency 35% with TensorRT quantization and distributed computing, while cutting training cycles 40% using synthetic data augmentation.
Machine Learning Engineer
Yahoo Beijing
Jun 2010 - Nov 2014 (4 years 5 months)
Developed deep neural networks for voice biometric authentication, achieving 99% precision and reducing unauthorized access by 40%, and built a CNN-based automated inspection pipeline to triple throughput. Implemented scalable MLOps pipelines with PyTorch and AWS SageMaker to reduce model retraining latency and improved authentication success rates by 15% YoY using novel acoustic feature extraction techniques.
Education
Degrees, certifications, and relevant coursework
Peking University
Master of Science, Computer Science – Artificial Intelligence
2007 - 2010
Earned an M.S. in Computer Science with a focus on Artificial Intelligence at Peking University (2007–2010).
Tech stack
Software and tools used professionally
Availability
Location
United StatesAuthorized to work in
Social media
Job categories
Skills
Interested in hiring Song?
You can contact Song and 90k+ other talented remote workers on Himalayas.
Message SongFind your dream job
Sign up now and join over 250,000+ remote workers who receive personalized job alerts, curated job matches, and more for free!
