# Engineering, Research at Greptile

> Onsite · United States · Unspecified · Software engineering · open. Source: https://alderhq.app/jobs/ashby%3Agreptile%3Ace47ec13-9aa2-4313-908d-8cfdb1e621be (read 2026-11-22).

## Facts (observed)

- Company: [Greptile](https://alderhq.app/companies/greptile)
- Location text: San Francisco
- Where it can be done from (inferred): Onsite · United States (evidence: "Location: San Francisco")
- Posted: 2026-08-22 (93 days ago)
- First seen by alder: 2026-11-02 · last verified: 2026-11-19
- Reposted: 0 times · edited: 1 times
- Apply on the employer's site: https://jobs.ashbyhq.com/greptile/ce47ec13-9aa2-4313-908d-8cfdb1e621be
- Skills listed (inferred): LLMs, Figma

## Pay and sponsorship

- Posted pay: $150k–$375k per year (from the listing's pay field)
- Market band for similar posted roles (SF Bay Area, n 2331): p25 $160k, median $200k, p75 $240k

## Description (the employer's text)

### Engineering, Research

Greptile is building the independent code validation layer for humans and AI agents. We combine code review, testing, and security with deep codebase context so teams can delegate more to coding agents and ship software they can trust.

Greptile reviews 5B lines of code every month for 22,000+ customers. A new repo starts using Greptile every 2 minutes. We also offer self-hosted, air-gapped deployments for enterprises across defense, financial services, and healthcare.

You'll push the quality of Greptile's agents and build the evals platform that tells us, and customers, whether a change actually made review better.

### Problems we're excited about

How do you measure whether a code review agent got better when ground truth is noisy and subjective?

Can sample-efficient RL or preference learning tune review behavior per customer?

Which multi-agent architectures, harnesses, and models catch the most real bugs per dollar?

### Trajectory

22,000+ customers

5B lines of code reviewed every month

Scaled from $0 to eight figures in ARR

Raised $80M from Menlo Ventures, Benchmark, Y Combinator, and Initialized

### Team

Small, talent-dense team who have scaled critical functions at companies like Open AI, Stripe, Google, and Figma.

### Responsibilities (in order of time spent)

Build and own the evals platform: benchmarks, datasets, graders, and regression tracking

Experiment with the latest agent and LLM techniques, and ship the ones that win

Design multi-agent systems, harnesses, and post-training experiments for code validation

Partner with product engineers to take prototypes to production

Stay current on research in LLMs, retrieval, and developer tooling

### Qualifications

B.S. Computer Science or equivalent

Research experience in CS, math, or physics

Strong programming skills and sharp product intuition

Bonus: ML/LLM/agent research, top-conference papers, post-training experience, your own evals or harnesses, or a research MS/PhD

### You'll like this role if

You're AI-enthusiastic, hack on side projects, and care about shipping research users feel

You want to work on a product that thousands of developers rely on

The chaos of high growth and things breaking is exciting to you

You like being in an office every day around other smart people who are excited about what they're building

You specifically prefer working in-person over working remote

---
Data: alder (https://alderhq.app). Cite https://alderhq.app/jobs/ashby%3Agreptile%3Ace47ec13-9aa2-4313-908d-8cfdb1e621be and the date you read it. Observed facts come from the employer's own listing; location reach, level, family and skills are inferences.
