AI agents · Product engineering
Cori · Interview coordination
I extended Cori for its launch: an AI agent that coordinates hiring, reports progress as it works, and waits for a recruiter’s approval before it touches a calendar.
Meet CoriEric Valens / Ventura
I’m a Senior Software Engineer at GoodTime.io and the founder of Chimpipi. I take ambiguous problems from research to launch, and build workflows that help local businesses get things done.
Selected work
Product engineering / Applied AIAI agents · Product engineering
I extended Cori for its launch: an AI agent that coordinates hiring, reports progress as it works, and waits for a recruiter’s approval before it touches a calendar.
Meet CoriAI evaluation · Data pipelines
I built and scaled the AI candidate evaluator, bringing together resume parsing, ATS data, and an in-house OCR pipeline. Personal identifiers are stripped before the prompt, inside an independently bias-audited screener.
My work at GoodTimeFounder · Full-stack engineering
I sit down with local business owners, find what is limiting them, and build the workflow. Websites, voice agents, and office automation, with training for the people who run it.
Explore ChimpipiNine years of shipping software. The last two have been AI features at GoodTime.io, working with the founder on the ambiguous projects: research the problem, find the right people, work out an approach, and see it through launch. That is how the product got its first generative AI, a resume evaluator inside an independently bias-audited screener, and Cori, the agent that schedules interviews and waits for a recruiter to say yes. I am careful about what AI does on someone's behalf. A person approves before anything is sent, published, or released. I am a huge fan of Grok Bot, and the film below is why. On the side I run Chimpipi: I sit down with local business owners, find what is limiting them, build the workflow, and train the people who run it.
I am a huge fan of Grok Bot. I pick the tickets, it researches each bug, and Cursor cloud agents write the fixes, one per bug, all at once. I test every one. Same hours, about ten times the work. A clean review is not a merge.
I pick the tickets. Grok Bot researches them and Cursor cloud agents fix them in parallel, several at once. That is the point. Same hours, about ten times the work. I do the testing. A clean review is not a merge. I designed the loop, then taught it to the team.
Grok Bot checks what I'm assigned, then reads the ticket, Slack, Notion, Drive, Datadog, the Loom, the screenshots, and the code. Each bug gets a brief. It does not reply in Slack.
Each brief goes to its own Cursor cloud agent. Four bugs get fixed side by side, at the same time.
CodeRabbit and SonarCloud review every PR. Then Claude Code and Codex review it as adversaries, trying to break the fix. The agents answer what gets flagged before I read it.
The most important step: I test every fix on stage, then prod.
A clean review is not a merge. I merge what passes and send back what doesn't. Grok Bot keeps notes on what worked, and the loop runs again.
On in the car, on a run, and between builds. None of them know it.
The timeline I read for AI news, and the people I look up to. All on X.