AI Product Engineer (Elixir)

About this role

Location: Remote from LatAm (preferred) Working hours: Full-time, aligned w/ Eastern or Central Time business hours (ET, CT). Job Type: Contractor, monthly payments in US dollars Benefits: 9 paid holidays and 10 days of PTO About Us WebRTC.ventures is one of the few software development agencies in the world dedicated exclusively to real-time applications. Originally founded as AgilityFeat in 2010, we began specializing in WebRTC in 2015.

We are headquartered in Charlottesville, VA, with a QA/testing center in Panama City, Panama, and a remote office in Bogotá, Colombia. With primary operations in North and South America, we serve clients around the globe. Our team has always been remote, which fuels our passion for real-time communications. About the role We are looking for an experienced Elixir developer who has shipped substantial production software and now directs AI coding agents for most implementation work, writing code directly where judgment requires it.

This is a hands-on, client-facing product engineering role. You will understand the problem, decide which parts agents should build and which need your direct attention, and own the delivery of reliable software. Fluent English is a must. B2+ or above is required for daily written and spoken client and team communication. What you’ll own Translate customer or market needs and high-level ideas into product decisions, clear specifications, acceptance criteria, and achievable weekly delivery goals.

Own architecture and implementation of Elixir-based applications, including integrations, APIs, data models, scalability, performance, security, and maintainability. Set up and operate autonomous coding-agent workflows: plan tasks, run parallel agents where useful, preserve context and state, recover failed or interrupted work, and manage execution cost and limits. Build a dependable delivery loop from specification/PRD to tracked issues, agent implementation, automated tests, CI, independent QA/security review, pull requests, and controlled merges and releases.

Use engineering judgment to challenge agent output, diagnose difficult problems, validate behavior end-to-end, and prevent regressions; write code directly when it is the best way to solve a problem. Communicate daily and proactively with the client and team: surface risks early, propose trade-offs and solutions, demonstrate working increments, and maintain visible progress without waiting for instructions. Continuously improve the agent harness, development standards, observability, and quality gates so the workflow becomes faster, safer, and less dependent on manual supervision.