Careers
Illustration: a person conducting an orchestra of small glowing cyan machines at workbenches in a dark navy workshop.

Careers Founding role

Agentic Software Engineer

Direct a fleet of AI coding agents across a real platform, and own everything they ship.

$185,000 base plus founding equity. Full time, remote in the US. Starts November 1, 2026. Reports to the founder and CEO.

Apply for this role

Run the software factory. Most of the code at Renkara is written by AI coding agents working in parallel, and it reaches production only through tests, reviews, and release gates we built ourselves. You would be the first engineer beside the founder: deciding what the agents build, writing the briefs and acceptance tests they work from, catching what they get wrong, and hardening the machinery that lets a very small team carry the work of an engineering department.

A day in the life

A composite day: the work is real, the timing is imagined, and no two days here look alike.

  1. Read what ran overnight

    Six agent lanes worked through the night. You read their reports before their code: four finished green, one is waiting on a decision, and one says a test is flaky.

    ___PRESERVE_BLOCK_4___

  2. Find the real bug

    The test isn't flaky. A study streak breaks when a learner's day crosses midnight in their own time zone. You write the regression test first, watch it fail, and only then fix the cause.

  3. Write the brief

    The best agent work starts with a short brief: the goal, the inputs, what done looks like, and one line of hard rules. You write three before lunch.

    ___PRESERVE_BLOCK_5___

  4. Lunch

    The agents keep working. You don't have to.

  5. Review like you mean it

    You read every diff from the morning lanes. One agent made a failing check pass by weakening the check. You reject it, explain why in one sentence, and the next attempt fixes the real problem.

  6. Judge the judges

    A model that grades generated practice questions is letting a class of defect through. You add a plain mechanical check that catches it first, so the model never has to.

  7. Certify the exact commit

    A release candidate goes through the certification gate: the exact commit, its full checks, and a receipt. When the founder gives the go, it ships.

    ___PRESERVE_BLOCK_6___

  8. Queue the night shift

    Briefs for the overnight lanes, each one small enough to finish and check. Then you close the laptop.

What you would own

  • Directing fleets of coding agents (Claude Code, OpenAI Codex, Gemini) across the AccelaStudy® AI platform: the adaptive learning engine, the API, the web app, and the content pipeline behind them.
  • The briefs, reviews, and acceptance tests that turn agent output into production code. Every defect fix ships with a regression test that fails if the bug comes back, and nothing merges on a red suite.
  • Our internal tools fleet, roughly thirty services covering tracking, deploys, observability, mail, analytics, and scheduling, and the agent tooling that runs on top of it.
  • Infrastructure on AWS through our own provisioning system, and releases through a certification gate that proves the exact commit before it reaches production.
  • Content quality at scale: the mechanical checks and model judges that decide whether generated course material is good enough to put in front of a learner.
  • Making the factory faster: finding where agents stall, waste work, or fail quietly, and fixing the cause instead of the symptom.

What we are looking for

  • Seven or more years writing production software, with real depth in Python and TypeScript.
  • Substantial hands-on experience directing AI coding agents on real codebases, not demos. Tell us what you shipped and how you kept it correct.
  • Judgment about where models are reliable and where they are not, and the testing discipline to prove which is which.
  • Comfort across the stack: backend services, React front ends, PostgreSQL, and AWS.
  • The habit of reading the diff. Agents write most of the code here, and you are accountable for all of it.

Helpful, not required

  • Evaluation harnesses, LLM-as-judge pipelines, or model-graded test suites.
  • Multi-agent orchestration, MCP servers, or developer tooling.
  • Education technology, adaptive learning, or psychometrics.
  • Being the first engineer at a company, or leading a small engineering team.

The terms

Base salary
$185,000
Equity
A founding-team grant, discussed in the offer conversation
Start date
November 1, 2026
Employment
Full-time , Exempt
Location
Remote (US)
Reports to
Founder & CEO

Where it goes. Structured to grow into Head of Engineering as the team is built underneath you. Milestone-gated, with the criteria written down before you start.

Benefits start on your first day. See everything working here includes.

Make your case

Email us whatever makes it best. A resume is fine. A short note about what you would do in your first ninety days is better.

Apply for Agentic Software Engineer

Or see the other open roles.