01
Execution reliability
Durable AI workflows
Designing and implementing the backend lifecycle around long-running AI work, with clear progress, recoverable state, and reliable access to completed results.
View work
Gyuho Lee / Software Engineer
Reliable systems.
Useful AI.
I build backend services and the infrastructure around AI. My work connects distributed systems, model serving, and careful evaluation.
Dallas–Fort Worth, Texas
Currently
Software Engineer at Viridia Solutions
Georgia Tech MS in CS · Expected Dec 2026
01 / Selected work
What I’ve been building.
02
Model serving
Self-hosted inference
Building LLM serving and evaluating it against complete application workloads, with attention to tail latency, throughput, and GPU resource use.
View work
03
Evaluation quality
Applied ML evaluation
Evaluating observational data with causal ML methods, careful feature validation, and explicit limits on what an estimate can establish.
View work
04
Data correctness
Backend reliability & data correctness
Improving event-driven services through production debugging, database tuning, and recovery tooling that makes operational intent explicit.
View work
Selected professional contributions, shared at a general level. Company implementation details are omitted.
02 / How I work
Build it. Understand it.
Make it dependable.
Think through failure.
Retries, interrupted work, and unexpected data are part of the problem. I design for those cases from the start.
Make evidence count.
I care about what a benchmark or model result actually supports, and the assumptions that could change the answer.
Build with people.
I mentor engineers, document decisions, and work through the interfaces between teams as carefully as the ones between services.
Off the critical path
A little room to play.
Small games, state machines, and the occasional high score.