Build evaluations to measure and improve AI reasoning in complex, high-stakes decisions.
Collaborate with engineers, product, and domain experts to turn observed failures into testable improvements.
Own the quality of shipped systems, from code reviews to monitoring and integration.
Elicit builds AI systems that help research and business teams make evidence-based decisions on high-stakes questions. They are a remote-friendly, collaborative team that values rigorous thinking, experimentation, and in-person retreats.
Design, build, and deploy AI agents that automate Mind Moves' internal operations, from contractor onboarding to market-rate research.
Decompose workflows into agent architectures—skills, tools, evals, memory, and orchestration—and build evaluation frameworks that measure and improve agent quality.
Own the full lifecycle from identifying workflows to prototyping, deployment, monitoring, and iteration, creating reusable infrastructure and documentation.
Mind Moves is a women-owned Washington, D.C.-based consultancy that helps businesses and governments design, build, and deploy human-in-the-loop AI solutions. It is a small, entrepreneurial team of AI pioneers that has delivered products serving millions of users and worked with Fortune 500 corporations and large government entities.
Design and deploy AI agents that automate sales, operations, and support workflows.
Own technical delivery end-to-end, from architecture to production and post-launch improvements.
Work directly with clients to understand workflows and explain technical tradeoffs.
Stello is an AI transformation firm that partners with mid-market and enterprise companies to automate repetitive work with AI agents. They work across sales, operations, and customer support, deploying solutions inside existing client tools.
Build production Agentforce agents including topics, actions, grounding, guardrails, and evaluation harnesses.
Implement actions across Flow, Apex, prompt templates, and external APIs, grounding answers in Data 360.
Design guardrails against prompt injection and data leakage, and run evaluation and regression tests before changes.
AspenView Technology Partners builds high-performing nearshore IT teams for North American clients, focusing on innovation and efficiency. They are a people-first, purpose-driven company that believes great culture drives great outcomes.
Own the AI coding agent's context and harness system design.
Own online and offline evals for the agent.
Leverage 100k+ users and 200k+ repos to guide experimentation.
Hercules builds an AI coding agent designed to make developers more productive. It is a startup with a founding team in San Francisco that values speed, excellence, and a bias for action.
Set the technical direction for AI engineering across the team: agent architecture patterns, evaluation methodology, deployment and monitoring strategies
Design the AI platform layer, including shared agent frameworks, tool integrations, and evaluation infrastructure
Work directly with clients on the most complex engagements, identifying new problem domains and ensuring production quality
Addepar is a global data and AI platform empowering investment professionals to turn complex financial information into actionable intelligence. With over 1,500 firms in 60 countries managing nearly $10 trillion in assets, Addepar fosters a culture of ownership, collaboration, and innovation.
Design and ship features end to end across data model, API, background workers, and UI.
Build and improve agent and model evaluation: sandboxed trials, scoring, and trustworthy aggregation of thousands of runs.
Improve benchmark task quality checks, integrate new models and agents, and keep the system fast and reliable as usage grows.
Turing accelerates superintelligence by providing high-quality datasets, reinforcement learning environments, and research benchmarks to frontier AI labs, and also deploys agentic AI systems for Fortune 500 enterprises. Headquartered in San Francisco, it operates with startup speed and an AI-forward culture.
Build tools for understanding traces, evaluating agent behavior, and turning results into improvements.
Run experiments on Mastra's primitives to find where they fall short and improve configuration and behavior.
Help turn findings into capabilities developers can use in their own applications.
Mastra is the open-source TypeScript framework for building AI agents. We're a small, fully remote team backed by Y Combinator and Spark Capital with $35M raised.
Own agent quality end-to-end for Bliro's chat and voice agents, including prompting, context, and model evaluation.
Build eval sets, LLM judges, and benchmarks to measure quality, latency, and cost with statistical rigor.
Establish a repeatable process for improving agent quality and set the long-term technology roadmap.
Bliro builds an AI assistant that handles desk work for field sales teams, with over one million customer touchpoints documented. Backed by leading investors and trusted by German Mittelstand companies, it's a small, high-performing team that values ownership and continuous improvement.
Study how engineers and customers use Archie in the field, identify where it succeeds or fails, and turn observations into actionable evaluations.
Recreate real-world engineering tasks and failure modes in repeatable environments for development teams.
Build and refine evaluation methodologies, including automated judges and human evaluation processes, to align with expert judgment.
P-1 AI is building Archie, an AI engineer agent for the physical world that works alongside human engineering teams. The company recently raised a $50 million Series A led by NEA and is driven by the mission of building superintelligence for engineering.
Set technical direction for production agent systems, including orchestration, tool use, memory management, and failure recovery.
Build and operate shared foundations: evaluation frameworks, observability, execution controls, and model access with routing and failover.
Work with security and platform teams to ensure scoped permissions, data isolation, audit trails, and safe execution of tools and code.
Cobalt runs AI agents in production that test real customer applications, operating at the forefront of the pentesting industry. The company is rapidly expanding, offering a collaborative culture with mentorship, competitive compensation, and a focus on growth.
Building multi-agent systems and implementing agent-based workflows and AI automation solutions.
Developing and optimizing prompt engineering strategies, including evaluation and iteration.
Integrating LLMs and AI models into cloud-based applications and APIs while researching new AI tools.
Software Mind develops impactful software solutions for global companies, partnering with tech giants and unicorns on transformative projects. The company builds cross-functional engineering teams and embraces a culture of openness, respect, grit, and cooperation.
Own quality for non-deterministic, multi-agent AI systems, building the evaluation harness that underpins AI governance.
Verify grounding and faithfulness, building automated checks to stop hallucinations from reaching production.
Red-team for prompt injection, jailbreaks, and tool misuse, and prove multi-tenant data isolation.
SOFICO Intelligence uses AI to extract value from back-office systems, building agentic AI for financial, leasing, fleet, and mobility data. Backed by global ERP leader SOFICO, it's a remote-first team of curious, self-directed builders.
Own the agentic development environment, ensuring AI agents can operate autonomously in cloud-based environments and execute full test suites.
Build MCP server integrations to connect agents to internal systems like CircleCI, Slack, Datadog, and GitHub.
Drive enablement by owning repo-wide agent documentation and developing tooling for engineers and non-engineers.
Hightouch is a data activation platform that helps companies sync customer data to their business tools. It is a remote-first startup with a fast-paced, high-ownership culture.
Interact with AI models across a wide range of prompts and topics.
Compare AI-generated responses side by side to assess quality.
Write clear, well-reasoned explanations supporting your evaluations.
Anyone AI creates high-quality training and evaluation data for frontier AI models. They work with remote independent contractors, focusing on expert human evaluation to improve AI accuracy and reliability.
Build AI-powered agents and tools end-to-end: identify problems with clients, prototype quickly, evaluate rigorously, and ship to production.
Design and implement LLM integrations including prompt engineering, function calling, structured outputs, and retrieval-augmented workflows.
Work directly with clients to understand workflows, validate AI behavior, and iterate toward production quality.
Addepar is a global data and AI platform empowering investment professionals to turn complex financial information into actionable intelligence. With over 1,500 firms in 60 countries and nearly $10 trillion in assets on the platform, Addepar fosters a culture of ownership, collaboration, and continuous learning.
Design and build automated evaluation pipelines for AI agents.
Develop testing frameworks for reasoning, multi-turn conversations, and adversarial scenarios.
Own the evaluation architecture and integrate observability tools in a fully remote team.
A company building production-grade AI agents and systems. They operate in a fully remote, globally distributed environment with a culture focused on innovation, autonomy, and continuous improvement.
Own the end-to-end product lifecycle of Joy.ai's AI agent reasoning layer, from discovery to launch.
Drive agent orchestration across skills, tools, and other agents within Asurion's ecosystem.
Define product-quality evals that prove the agent is useful, trustworthy, and ready for real customers.
Asurion is a global tech care leader providing protection, repair, installation, and expert support for consumer devices. Its products and services help nearly 300 million customers worldwide, with a culture that values customer-first, teamwork, and integrity.
Build AI features inside real clinical workflows, like checking notes against payer and compliance requirements before finalization.
Design evals to prove those features are right and keep them honest as models change.
Prototype AI capabilities, including voice agents, and take the best from prototype to production.
Ease builds the AI-native operating system for behavioral health, replacing disconnected tools with one platform for CRM, EHR, and revenue cycle management. Backed by a $41M Series A, this rapidly growing NYC team has no product managers—engineers talk to customers and ship in days.