Anthony Ledesma, Staff AI Engineer

Now: Staff AI Engineer at TaskRay, Jun 2026 to now. Agent systems, and the evals that keep them honest.

1. 2016GoDaddy, Tier 3 Support, then Software Engineer I Nov 2016 to Oct 2019

   Started in enterprise hosting support and taught myself to code. Shipped React and Node features for apps serving 20M+ customers.

2. 2019GoDaddy, Software Engineer Oct 2019 to Apr 2024

   Smart Templates inference cost too much, so I refactored the retrieval and generation pipeline and moved it to a cheaper model. Inference cost fell 93% and latency 71%. Mentored junior engineers on prompt engineering and AI integration.

3. 2024GoDaddy, Senior Software Engineer Apr 2024 to Apr 2026

   Built MCP-style agent tool layers before the protocol was standardized, hand-rolling the connectors so no vendor SDK sat waiting on security review. Shipped six agent layers in one conversation for a high-visibility executive demo, each tool gated on conversation state. Worked across 5 engineering groups to codify and share AI patterns, and cut CI/CD runtimes 83%.

4. 2025McLean Forrester LLC (Contract), Senior AI Engineer Sep 2025 to Mar 2026

   Gated releases on a tolerance band over N runs, so a release can't flip on noise alone. When an upstream change dropped quality 39 points against a baseline spread of about 2, the gate caught it before it shipped. Took over an inherited Terraform migration on GCP and delivered full platform parity, with zero secrets in state.

5. 2026Tracine, Open-source tooling for AI coding agents 2026

   Tools I built for my own agents, then open-sourced. guard allows, denies or asks on agent tool calls before they run. convo makes every session and tool call searchable. eval-kit says whether an eval got better or it's noise, and how many runs it takes to tell.

6. NOW TaskRay, Staff AI Engineer Jun 2026 to now

   LLM and agentic systems.

   GitHubLinkedIn