Back home
Beacon HealthYC W262026 — Present

I take customer problems all the way to production.

As a Forward Deployed Engineer, I own customer work from the first conversation through workflow design, agent development, rollout, and day-to-day reliability. The work crosses clinical systems, customer operations, and core engineering — and I build the tools and practices that let the rest of the team move faster too.

Visit Beacon Health ↗AI workforce for healthcare
Beacon Health website showing its AI workforce for healthcare and agent demonstration.
Beacon Health's live website · beaconhealth.ai

I own the full deployment.

A customer often arrives with a painful process and no clear technical specification. I work directly with them to understand the job, find the path through their systems, build the agent, validate its decisions, launch it safely, and own what happens after it goes live.

  1. 01Discover

    Translate open-ended needs into a workflow that can actually run.

  2. 02Build

    Engineer agents and internal tools around real clinical systems.

  3. 03Operate

    Monitor production, resolve failures, and improve cost and reliability.

  4. 04Scale the method

    Turn working deployments into reusable practices for the team.

9customer workflows owned
~15Kproduction agent runs
98%success rate

Shipping the work and improving the system behind it.

01 / Build instead of buy

Replaced a vendor with in-house capability in one week.

I built an internal AI tool that took over a production workflow from an external automation vendor. That single workflow cost 91% less to run, saving about $5K each month.

02 / From ambiguity to launch

Turned a loose customer request into a live agent.

With a healthcare platform serving more than 20,000 providers, I moved from an initial, open-ended conversation to a deployed workflow in under a week.

03 / Make the next deployment easier

Built a repeatable way to ship reliable agents.

I distilled my approach to exploring unfamiliar systems, validating behavior, and rolling out live writes into a shared workflow-building method used by the engineering team.