very early LinkedIn, Twitter, Allen Institute for AI, usually relevance, recommenders, recsys, model infra, e2e model-to-product development. Enterprise search startup (field CTO-ish role). Then on to large enterprise roles, where I spend a lot of time with lawyers, the business side of the house, external partner integration, build/buy decisions, general purpose "scheming vizier"-roles. Still an IC, if you can believe it.

AI Governance layer components (protocol enhancements, registries, gateways, policy enforcement, workflows), Alignment research (abliteration, model introspection, red-teaming, weaker-model supervision), Harness engineering, Stateful agents.

multi-agent alignment research, virtual tools infrastructure, e2e harness engineering experimentation

Liam Thompson is a Senior Software Engineer at IO Intelligence, where he leads production architecture, migrates services from AWS ECS to EKS, and builds core IoT data platforms. He is also involved in shipping generative AI features and ETL pipelines using OpenSearch, Snowflake, and Ray on Kubernetes. Through Imminent Latency, Liam architects AI/ML pipelines and cloud-native microservices for early-stage startups. Previously, he developed serverless SaaS architectures using AWS Amplify, Lambda, and DynamoDB, and served as a TA for an agentic AI capstone. Liam has 6+ years of experience and is looking for community, friendships, co-founders, investors, and speaking opportunities. He is open to introductions and prefers to be contacted via email.

Cloud-native microservices, IoT data platforms, AI/ML systems, generative AI, agentic AI, Kubernetes (EKS), serverless architectures, ETL pipelines, OpenSearch, Snowflake, Ray, technical architecture, startup advisory, cloud-native infrastructure, model serving.

Leading IO Intelligence’s production architecture, migrating services from AWS ECS to EKS, and building core IoT data platforms. Shipping generative AI features and ETL pipelines using OpenSearch, Snowflake, and Ray on Kubernetes. Through Imminent Latency, architecting AI/ML pipelines and cloud-native microservices for early-stage startups. Previously developed serverless SaaS architectures using AWS Amplify, Lambda, and DynamoDB, and served as a TA for an agentic AI capstone.