Open to opportunities

Frank Cheng

@frankcheng

Message

AI/ML and agentic systems engineer building low-latency, reliable decisioning platforms at massive scale.

United States

Message

What I'm looking for

I’m looking to build agentic AI and LLM-driven decisioning systems that stay reliable and observable in production, delivering low-latency, real-time impact at scale—ideally with strong ownership across distributed inference, RAG, and data pipelines.

I’m an AI Engineer with 10+ years of experience specializing in AI/ML systems, agentic AI, and distributed architectures. I’ve built real-time decisioning platforms, LLM-powered applications, and large-scale data pipelines with a strong focus on reliability, observability, and business impact.

At The Trade Desk, I architected KokaiAI decisioning features and designed distributed inference pipelines processing 13M+ impressions per second, improving bid optimization accuracy and campaign ROI by ~18%. I also built agentic AI trading workflows with autonomous budget allocation and bid adjustments (human override controls), improving optimization efficiency by 25%, and delivered an LLM-powered campaign insights engine with OpenAI and LangGraph that reduced analysis time by 40%.

I engineered real-time inference services in Python (FastAPI) with sub-50ms latency while handling millions of scoring requests per second. Previously at Adobe, I built AdobeSensei-powered decisioning and scalable ML pipelines (Python, Spark, AWS) across billions of daily events, plus streaming and batch ingestion frameworks that improved data availability SLAs by 35%.

Experience

Work history, roles, and key accomplishments

Current

Senior Software Engineer

Current

The Trade Desk

Jan 2023 - Present (3 years 5 months)

Architected KokaiAI decisioning features and distributed inference pipelines processing 13M+ impressions/second, improving bid optimization accuracy and campaign ROI by ~18%. Built agentic AI trading workflows and an LLM-powered campaign insights engine, reducing analysis time by 40% and achieving sub-50ms inference latency at millions of scoring requests/second.

Python TypeScript fastAPI LLM RAG Kafka Flink AWS Kubernetes

AI/ML Engineer

Adobe

Sep 2018 - Dec 2022 (4 years 3 months)

Engineered core Adobe Experience Platform services unifying real-time customer profiles across billions of daily events and delivering AdobeSensei-powered personalization and recommendations. Built scalable ML and streaming/batch ingestion pipelines, improving data availability SLAs by 35% and reducing experience query data retrieval latency by 40%.

Python Apache Spark AWS Kafka REST APIs Identity Resolution Machine Learning Pipelines Batch Ingestion Pipelines Kubernetes Adobe Experience Platform

Software Engineer

Deloitte

Jun 2016 - Jul 2018 (2 years 1 month)

Delivered full-stack applications for enterprise and government clients using React, Java, Spring Boot, and Node.js, standardizing API integrations to improve reliability. Implemented AWS cloud-native solutions with CI/CD, cutting deployment time by 50%, and built identity/session management to reduce duplicate user records.