For large-model training and evaluation scenarios, participate in formulating and continuously iterating data quality standards, judgment rules, and acceptance criteria.
Lead BaaS platform technology strategy and architecture by defining the technology strategy and roadmap, ensuring architecture aligns with Group Technology and enterprise standards, and acting as the design authority for BaaS solutions.
Design, build, and maintain MLOps pipelines for training, testing, deploying, and monitoring AI/ML models in production, ensuring stable, reliable, and scalable performance.
Design and build LLM-powered applications using models such as Claude, GPT-4, Gemini, or open-source equivalents (Llama, Mistral), selecting the right model based on latency, cost, and quality tradeoffs.