AI evaluation startups hiring now

Dev Tools AI

San Francisco, US · Pre Seed

Dev Tools AI assist in the rapid development of web and mobile app tests. It offers an AI-powered automation service that uses AI to automate testing of web and mobile applications without requiring access to page source code.

36 open roles21% remote

HackerRank

Mountain View, US · Series D · YC S11

HackerRank is a technology hiring platform that is the standard for assessing developer skills for 2500+ companies around the world.

27 open roles21% remote

Prolific

London, GB · Series A · YC S19

Prolific connects researchers, AI developers and product teams with real people who take part in studies online. It is built for collecting high-quality, human data quickly and reliably, from training AI models to running behavioral research.

26 open roles94% remote

Surge AI

San Francisco, US · Bootstrapped

Surge AI is a data labeling platform that connects AI developers with a curated network of expert human annotators. It specializes in providing high-quality data for training, evaluating, and aligning advanced AI systems, particularly Large Language Models.

19 open roles100% remote

Goodfire

San Francisco, US · Series B

Goodfire is an AI research lab using interpretability to turn AI into something that can be understood, debugged, and shaped like software. Its product, Ember, is a model-agnostic interpretability platform which decodes the neurons inside of an AI model to give direct, programmable access to its internal processes.

18 open roles

Immersive

Bristol, GB · Series C

Immersive helps organizations prove and improve their cyber resilience by assessing people, teams, agents, and AI-enabled workflows under pressure.

17 open roles94% remote

Artificial Analysis, Inc.

San Francisco, US

Artificial Analysis is an independent AI benchmarking and analytics company. It evaluates and compares artificial intelligence models and providers.

16 open roles0% remote

Arena

San Francisco, US · Series A

Arena.ai is an AI evaluation platform that helps frontier AI labs and enterprises measure, benchmark, and improve AI models using real-world human feedback. Arena is an AI evaluation platform that benchmarks AI models and agents using real-world human feedback The company operates one of the world's largest AI evaluation platforms, powering public leaderboards and enterprise products for evaluating language models, multimodal systems, and AI agents.

15 open roles13% remote

Pangram Labs

Brooklyn, US · Seed

Pangram Labs provides technology solutions designed to detect AI-generated content at scale. The system functions by attempting to determine what a passage sounds like an LLM or a human.

14 open roles0% remote

Gray Swan AI

Pittsburgh, US · Series A

Gray Swan AI is an AI security company that develops tools for evaluating and protecting large language model deployments.

13 open roles77% remote

White Circle

Dover, US

White Circle is an AI Safety company building the safety, reliability, and optimization layer for AI systems. Its platform uses natural-language policies to automatically test, enforce, and continuously improve AI model behavior at scale.

11 open roles64% remote

Unusual

San Francisco, US · Seed · YC F24

Unusual helps brands in the Fortune 100 and scale-ups like Change.org and Astronomer drive more sales and win more market share by shifting how AI models perceive their offerings. Unusual creates and suggests media engineered to improve how AI talks about your brand – driving you more leads. Unusual helps you change AI's mind about your brand to get more leads.

9 open roles11% remote

Dynamo AI

San Francisco, US · Series A · YC W22

Dynamo AI delivers auditable AI guardrails, hallucination checks, red-teaming, and observability so enterprises can productionize AI with confidence.

9 open roles0% remote

AIUC

San Francisco, US · Seed

AIUC certifies and insures AI agents to unlock secure adoption. AI companies certify against AIUC-1 to earn enterprise confidence.

7 open roles0% remote

Ooak Data

San Francisco, US · YC S26

We turn company data into training data

7 open roles0% remote

Crosschq

Danville, US · Series A

Crosschq is an AI-powered Hiring Intelligence platform that helps companies improve hiring outcomes and build high-performing teams. Its suite of applications includes recruiting automation, predictive screening, fraud detection, interview intelligence, and advanced hiring analytics.

7 open roles17% remote

Workhelix

San Francisco, US

Workhelix provides software that helps companies find opportunities to use AI, track adoption, and measure the resulting return on investment.

7 open roles

Hyperbound

San Francisco, US · Series A · YC S23

The platform analyzes sales calls to identify winning behaviors and transforms those insights into customizable AI buyer roleplays that drive behavior change at scale.

7 open roles0% remote

Irregular

Tel Aviv, IL · Series A

Irregular is a frontier security lab that builds next-generation defenses through high-fidelity research platforms that simulate and monitor real-world AI security scenarios. It serves as a trusted partner to frontier AI companies like OpenAI, Anthropic, and Google DeepMind.

7 open roles0% remote

Vals AI

San Francisco, US · Seed

Vals AI assesses model performance on domain-specific tasks in finance, law, software, healthcare, and more.

7 open roles0% remote

Dome Systems

Menlo Park, US · Seed

The platform connects to systems to identify every agent, evaluates every action against policy before execution, and records every governed decision for auditing.

6 open roles0% remote

Idler

San Francisco, US · Seed · YC S25

Idler builds reinforcement learning environments that encode real human expertise into frontier AI systems.

6 open roles0% remote

METR

berkeley, US

METR is a research nonprofit that evaluates whether advanced AI systems can cause catastrophic risks. It develops evaluation methods and software to assess these systems.

6 open roles0% remote

Patronus AI

San Francisco, US · Series A

Patronus AI is a frontier lab training the first Digital World Models, which predict and simulate agent actions in digital workflows. The company also provides an automated evaluation and security platform for LLMs, enabling enterprise development teams to score LLM performance and detect mistakes at scale.

5 open roles75% remote

Mechanize, Inc.

San Francisco, US · Seed

Mechanize builds reinforcement learning environments that frontier AI labs use to train and evaluate their coding models.

4 open roles0% remote

Hamming AI

San Francisco, US · Seed · YC S24

Hamming automates QA for voice AI agents. With one click, thousands of our agents call our customers’ agents across accents, background noise, and personalities—then we generate crisp bug reports and production-grade analytics.

4 open roles

GPTZero

New York City, US · Series A

GPTZero develops an AI detection platform that determines whether text was generated by AI or written by a human. The tool measures perplexity and burstiness in text. For developers, GPTZero offers tools to scan training datasets and filter out AI-generated content.

4 open roles0% remote

Preference Model

San Francisco, US

Existing frontier models are brittle when applied to real-world ML tasks, so the company builds RL environments that reflect real-world complexity with diverse tasks and robust reward functions.

4 open roles0% remote

Cartpole

San Francisco, US · Seed

Cartpole creates reinforcement learning environments for training frontier models.

4 open roles0% remote

Armilla AI

Toronto, CA · Seed

Armilla AI is a Lloyd's Coverholder and MGA providing purpose-built AI liability insurance and AI software performance warranties for generative AI and AI agents. The company assesses and quantifies the risk level of AI models to provide AI verification and warranties backed by leading insurers.

3 open roles