Arjhine Ty
About

AI Systems Engineer

I design AI systems that behave credibly under real use. Most of my work sits in the unglamorous layers: evaluation, latency budgets, memory, and the instrumentation that tells you whether a demo is a product. Since June 2026 I have been Chief AI Officer at **Agylion**, where I own the agent platform behind RepReady, a sales roleplay system used by enterprise teams. Before that I built PocketPatient, a clinical interview simulator used in medical education, and spent two years as Senior AI Engineer at Northstar Health, a Manila healthtech company. I write about evaluation methodology, agent memory, and latency — usually with numbers attached. Everything on this site links to a source.

Skills

What I work with

AI systems

LLM evaluationRAGAgent orchestrationFine-tuning

Engineering

TypeScriptNext.jsPostgres / SupabaseDocker

Research practice

Knowledge distillationAblationsBenchmarking
Experience

Where I've worked

Chief AI Officer

2026 — present

Agylion

Own the agent platform: platform architecture, the evaluation layer, and the team that ships RepReady.

Senior AI Engineer

2024 — 2026

Northstar Health

Clinical NLP and the ML platform behind a Manila healthtech product used in 40 clinics.

ML Engineer

2022 — 2024

DataSpring Labs

First ML role: model training, data pipelines, and the unglamorous work of making eval sets that mean something.

BS Computer Science

2018 — 2022

University of the Philippines Diliman

Thesis on retrieval for code-switched clinical text — the seed of everything built since.