Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

I'm Ian L Juch, a Staff AI Systems Engineer with 10+ years building production LLM platforms and distributed systems at Adobe, Cohere, and Netflix. I design systems that operate under real constraints like latency, cost, correctness, and reliability across retrieval, agent orchestration, and evaluation pipelines. I own systems end-to-end from data ingestion to deployment, ensuring observability and debuggability in production. My work spans backend orchestration powering generative workflows, enterprise LLM retrieval infrastructure, and large-scale experimentation and traffic control to keep models safe and performant.…I'm Ian L Juch, a Staff AI Systems Engineer with 10+ years building production LLM platforms and distributed systems at Adobe, Cohere, and Netflix. I design systems that operate under real constraints like latency, cost, correctness, and reliability across retrieval, agent orchestration, and evaluation pipelines. I own systems end-to-end from data ingestion to deployment, ensuring observability and debuggability in production. My work spans backend orchestration powering generative workflows, enterprise LLM retrieval infrastructure, and large-scale experimentation and traffic control to keep models safe and performant.

Ian L Juch

AI Engineer, AI Developer, Data Scientist, +2





I'm Ian L Juch, a Staff AI Systems Engineer with 10+ years building production LLM platforms and distributed systems at Adobe, Cohere, and Netflix. I design systems that operate under real constraints like latency, cost, correctness, and reliability across retrieval, agent orchestration, and evaluation pipelines. I own systems end-to-end from data ingestion to deployment, ensuring observability and debuggability in production. My work spans backend orchestration powering generative workflows, enterprise LLM retrieval infrastructure, and large-scale experimentation and traffic control to keep models safe and performant.…I'm Ian L Juch, a Staff AI Systems Engineer with 10+ years building production LLM platforms and distributed systems at Adobe, Cohere, and Netflix. I design systems that operate under real constraints like latency, cost, correctness, and reliability across retrieval, agent orchestration, and evaluation pipelines. I own systems end-to-end from data ingestion to deployment, ensuring observability and debuggability in production. My work spans backend orchestration powering generative workflows, enterprise LLM retrieval infrastructure, and large-scale experimentation and traffic control to keep models safe and performant.

Available to hire

I’m Ian L Juch, a Staff AI Systems Engineer with 10+ years building production LLM platforms and distributed systems at Adobe, Cohere, and Netflix. I design systems that operate under real constraints like latency, cost, correctness, and reliability across retrieval, agent orchestration, and evaluation pipelines.

I own systems end-to-end from data ingestion to deployment, ensuring observability and debuggability in production. My work spans backend orchestration powering generative workflows, enterprise LLM retrieval infrastructure, and large-scale experimentation and traffic control to keep models safe and performant.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Intermediate

Intermediate

Intermediate

Intermediate

Intermediate

Intermediate

Work Experience

Staff AI Systems Engineer at Adobe

January 1, 2023 - Present

Built backend orchestration systems powering Firefly generative workflows, focusing on how models operate inside real creative pipelines (asset generation, editing, transformation). Designed a multi-stage execution pipeline spanning retrieval, prompt construction, model inference, and post-processing with explicit state tracking. Implemented context-grounding to inject user assets, project state, and prior actions, improving consistency and reducing stateless prompt failures. Created a tool-execution layer enabling models to call deterministic services (asset lookup, transformations, validation) to reduce reliance on pure generative reasoning. Devised model rollout and gating strategies balancing latency, cost, and content safety. Built evaluation constructs targeting grounding errors and missing context, and implemented full request-level observability tracing across retrieval → reasoning → generation to debug multi-step agent behavior.

Senior Applied AI / Platform Engineer at Cohere

May 1, 2020 - December 1, 2022

Built developer-facing platform powering generation and embedding APIs for enterprise use. Designed retrieval pipelines combining embeddings, vector search, and reranking to support grounded generation and reduce hallucinations. Implemented hybrid retrieval (dense + lexical) to improve recall on long-tail and entity-heavy queries. Built structured orchestration pipelines (retrieve → rank → prompt construction → generate → validate) with deterministic control over multi-step workflows. Addressed production inference constraints including batching, request queuing, and latency variability to stabilize API performance. Developed evaluation pipelines linking retrieval quality, prompt construction, and output correctness. Created internal debugging tooling exposing retrieval misses, ranking failures, and prompt construction issues.

Senior Applied AI Engineer at Netflix

June 1, 2019 - April 1, 2020

Built backend services within Netflix Experimentation Platform, enabling safe evaluation of product changes across global user traffic. Designed traffic routing systems controlling experiment exposure while maintaining statistical validity under uneven traffic distributions. Implemented staged rollout mechanisms (progressive ramp-up, kill switches, fallback paths) preventing faulty experiments from impacting large user segments. Built APIs enabling dynamic configuration of experiments without requiring redeployments. Improved observability enabling real-time detection of anomalies, regressions, and misconfigured experiments.

Software Engineer at Netflix

July 1, 2016 - May 1, 2019

Built distributed services supporting Traffic Routing & Experimentation APIs, handling high-throughput request routing across global regions. Developed internal platform components enabling safe feature deployment tightly integrated with experimentation workflows. Designed low-latency request routing systems ensuring consistent performance under globally distributed traffic. Implemented observability improvements (metrics, logging, tracing), reducing time to detect and resolve production issues. Built resilience mechanisms including retries, circuit breakers, and graceful degradation for high-availability services. Participated in on-call rotations for critical infrastructure, diagnosing failures across distributed systems.

Software Engineering Intern at Netflix

June 1, 2015 - August 1, 2015

Contributed backend code to services supporting Streaming Platform Infrastructure, focusing on reliability and service coordination. Built internal tooling improving deployment workflows and service monitoring. Improved logging and instrumentation enabling better debugging of distributed service interactions.