careers

Design and build the distributed systems that power the aligned collective. Work across systems software, infrastructure, and services, with ownership from architecture through production operation.

What you’ll do

• Design distributed services for storage, messaging, and scheduling that remain reliable under high load and partial failure.
• Make explicit tradeoffs in consistency, availability, performance, security, and cost as systems scale.
• Lead technical designs and migrations across components and teams, and mentor engineers through implementation and review.
• Profile production systems, resolve complex failures, and improve reliability through observability, testing, and capacity planning.

What you’ll bring

• Demonstrated experience designing, building, and operating very large-scale distributed systems, with ownership of architectural decisions and production outcomes.
• Strong programming skills across multiple languages such as Python, Swift, Rust, and TypeScript, with depth in systems programming.
• Strong foundations in data structures, algorithms, operating systems, networking, concurrency, and memory management.
• Practical understanding of replication, partitioning, consistency, fault tolerance, backpressure, and failure recovery.
• Experience profiling and debugging complex production systems, including CPU, memory, storage, network I/O, and tail latency.
• Experience shipping and evolving reliable software through automated testing, observability, safe deployments, and backwards-compatible migrations.
• Technical leadership across teams: defining architecture, writing clear design documents, mentoring engineers, and turning ambiguous goals into shipped systems.

Compensation

We offer a base salary of $200,000–$800,000 and a meaningful equity grant, depending on experience and background, along with competitive benefits.

We believe agents are the most disruptive technology of this decade. As we move from the chat era to the agentic era, the interface is critical to how people work with agents and the collective. We believe the best interface is no interface. Your job is to figure out how to achieve this by designing and building experiences that make complex capabilities accessible with as little effort as possible.

What you’ll do

• Design and build interfaces for human-agent and human-collective interaction across native and web applications.
• Prototype experiences that use context, voice, direct manipulation, and existing workflows to reduce the need for prompts and screens.
• Make the collective’s work understandable and accessible, revealing controls and detail when they help people make a decision.
• Take ideas from working prototypes to polished, reliable products, and test whether they reduce effort for real users.

What you’ll bring

Experience designing and shipping interactive software, with strong engineering skills in Swift and SwiftUI, React and TypeScript, or similar tools. Bring working prototypes and products that show your understanding of human-computer interaction and your ability to simplify complex workflows.

Compensation

We offer a base salary of $200,000–$800,000 and a meaningful equity grant, depending on experience and background, along with competitive benefits.

We believe agents are the most disruptive technology of this decade. Moving from the chat era to the agentic era requires rethinking how people and the collective work together. We believe the best interface is no interface. Your job is to discover and build interactions that let people express intent, delegate work, and guide agents naturally, with less prompting and coordination.

What you’ll do

• Design and implement interaction models for expressing goals, sharing context, and coordinating work with agents and the collective.
• Explore when agents should act on context, ask for clarification, surface progress, or hand work back to a person.
• Build ways to correct, interrupt, and redirect the collective without requiring people to manage individual agents.
• Run experiments with real users to measure understanding, effort, trust, and task outcomes, then turn findings into product changes.

What you’ll bring

Experience designing and building interactive systems, with strong prototyping and software engineering skills. Bring an understanding of human-computer interaction, experience testing ideas with users, and the ability to connect interface decisions to agent behavior.

Compensation

We offer a base salary of $200,000–$800,000 and a meaningful equity grant, depending on experience and background, along with competitive benefits.

Help people discover and adopt Agentastic by solving concrete problems with the product. Combine engineering and customer understanding to build useful demonstrations, integrations, and paths from first use to lasting value.

What you’ll do

• Work with prospective users to understand their workflows and build working examples around them.
• Develop integrations, onboarding tools, and technical content that help users get started.
• Measure activation and adoption, run focused experiments, and bring product feedback to the team.

What you’ll bring

An engineering background with experience in product growth, solutions engineering, or developer-facing work. You should be comfortable building integrations, explaining technical ideas, and using data alongside customer conversations.

Compensation

We offer a base salary of $200,000–$800,000 and a meaningful equity grant, depending on experience and background, along with competitive benefits.

Define the design of Agentastic and how people work with agents. Own the path from understanding a user’s task to an interaction that feels clear, useful, and consistent across the product.

What you’ll do

• Research workflows and prototype new ways to delegate, supervise, and collaborate with agents.
• Create interaction flows, visual designs, and a practical design system for our applications.
• Work closely with engineers through implementation and refine the product using real user feedback.

What you’ll bring

A portfolio showing thoughtful interaction and visual design for complex software. Bring experience with prototyping, user research, and shipping products, plus the ability to explain your decisions and work through technical constraints.

Compensation

We offer a base salary of $200,000–$800,000 and a meaningful equity grant, depending on experience and background, along with competitive benefits.

Build the runtime that turns models into dependable agents. Own how agents use context and tools, maintain progress across sessions, and recover when work fails.

What you’ll do

• Develop execution loops, tool interfaces, and context management for tasks that span multiple sessions.
• Build checkpointing, interruption, and recovery so agents can resume work without losing progress.
• Integrate sandboxing and permissions, and use evaluations to improve execution quality, latency, and cost.

What you’ll bring

Experience building runtimes, developer tools, or agent systems. Bring strong debugging skills, familiarity with model tool use, and a practical understanding of concurrency, process isolation, and persistent state.

Compensation

We offer a base salary of $200,000–$800,000 and a meaningful equity grant, depending on experience and background, along with competitive benefits.

Define and measure whether agents do useful work and follow user intent. Build evaluations that exercise complete workflows and make improvements and regressions visible across the whole agent system.

What you’ll do

• Turn real workflows and failure cases into reproducible tasks with clear success criteria.
• Build isolated evaluation runs and graders that check outcomes and traces against human judgment.
• Compare model and harness changes across task completion, instruction following, reliability, and cost.

What you’ll bring

Experience with Python, experimental design, and evaluating models or complex software. Bring careful judgment about measurement, statistical variation, and grader quality, plus the ability to turn failures into useful tests.

Compensation

We offer a base salary of $200,000–$800,000 and a meaningful equity grant, depending on experience and background, along with competitive benefits.

Build systems that let specialized agents work together on shared goals. Make delegation, communication, and shared state reliable as work moves between agents.

What you’ll do

• Design task decomposition, agent selection, communication, and handoffs for collaborative workflows.
• Build shared context and coordination mechanisms that resolve conflicting work and avoid duplicated effort.
• Evaluate when collaboration improves results and tune coordination for quality, reliability, and resource use.

What you’ll bring

Experience with distributed systems, asynchronous workflows, or agent orchestration. Bring strong systems engineering skills and an experimental approach to understanding how coordination choices affect the quality of collective work.

Compensation

We offer a base salary of $200,000–$800,000 and a meaningful equity grant, depending on experience and background, along with competitive benefits.

Develop methods that keep agents and the collective aligned with human intent as their capabilities grow. Work across ML research and engineering to study how agents behave, learn from human feedback, and remain understandable and steerable during complex tasks.

What you’ll do

• Turn questions about agent behavior into testable hypotheses, controlled experiments, and reproducible findings.
• Design evaluations that measure intent following and distinguish real improvements from reward exploitation or misleading reports.
• Develop and test scalable methods for people to supervise, correct, and redirect agents working together.
• Study coordination failures, goal drift, and robustness across long tasks, changing contexts, and adversarial conditions.
• Implement and evaluate alignment interventions using human feedback, preference learning, or reinforcement learning.

What you’ll bring

• Strong Python and PyTorch skills, with experience building and debugging ML training or evaluation systems.
• Experience designing controlled ML experiments, choosing meaningful baselines, analyzing results, and explaining the limits of the evidence.
• Familiarity with LLM post-training, reinforcement learning, or preference learning, supported by hands-on research or engineering work.
• Experience building behavioral evaluations and identifying when models exploit a scoring function instead of following human intent.
• An understanding of alignment research and how oversight, feedback, and correction can work across multiple agents.
• Ability to own an ambiguous research problem from implementation and experimentation through reproducible results and clear technical communication.

Compensation

We offer a base salary of $200,000–$800,000 and a meaningful equity grant, depending on experience and background, along with competitive benefits.

Build datasets that teach agents useful behavior and support reliable research. Own data quality from source selection and generation through training-ready examples and evaluation splits.

What you’ll do

• Curate and generate supervised examples, agent trajectories, and preference feedback for post-training.
• Build versioned pipelines for validation, deduplication, labeling, and reproducible dataset releases.
• Track provenance and source permissions, protect sensitive data, and keep training and evaluation data separate.

What you’ll bring

Experience with Python, data pipelines, and datasets for machine learning. Bring careful judgment about source quality, annotation, and evaluation leakage, plus the ability to investigate how data changes affect model behavior.

Compensation

We offer a base salary of $200,000–$800,000 and a meaningful equity grant, depending on experience and background, along with competitive benefits.

Improve how models reason, use tools, and complete agent tasks through reinforcement learning. Develop training approaches and experiments that produce measurable improvements in useful behavior.

What you’ll do

• Experiment with training objectives, task mixtures, and feedback for work that takes many steps.
• Investigate unstable training and rewards that favor shortcuts over correct outcomes.
• Compare trained models on held-out tasks and make successful experiments reproducible.

What you’ll bring

Practical experience with model training, Python, and an ML framework such as PyTorch. Bring strong experimental judgment and the engineering skills to investigate both training behavior and data quality.

Compensation

We offer a base salary of $200,000–$800,000 and a meaningful equity grant, depending on experience and background, along with competitive benefits.

Build the infrastructure that makes reinforcement learning experiments reliable and efficient. Help researchers run, understand, and repeat training jobs as workloads grow.

What you’ll do

• Build distributed services for collecting agent trajectories, scheduling training, and managing checkpoints.
• Profile GPU workloads and data movement to improve throughput and resource use.
• Make experiment state, failures, and performance visible so interrupted work can be recovered.

What you’ll bring

Experience operating distributed systems or ML infrastructure. Familiarity with GPU workloads, training frameworks, and performance profiling is valuable, along with a disciplined approach to reliability and debugging.

Compensation

We offer a base salary of $200,000–$800,000 and a meaningful equity grant, depending on experience and background, along with competitive benefits.

Build the tasks and environments agents learn from. Create realistic, reproducible settings where progress can be measured and success reflects useful work.

What you’ll do

• Develop interactive tasks with tools, persistent state, and clear outcome checks.
• Build isolated execution and reset mechanisms for repeatable training and evaluation.
• Analyze agent traces to find task flaws, scoring exploits, and gaps in coverage.

What you’ll bring

Experience building execution environments, simulations, test infrastructure, or data pipelines. You should be comfortable with Python, sandboxing, and defining what correct behavior means in complex tasks.

Compensation

We offer a base salary of $200,000–$800,000 and a meaningful equity grant, depending on experience and background, along with competitive benefits.

Build agents that help improve the systems used to develop agents. Turn parts of the research and engineering process into repeatable workflows whose results can be independently checked.

What you’ll do

• Create agent workflows for experiment design, implementation, debugging, and analysis.
• Connect proposed changes to controlled experiments and independent evaluations.
• Measure improvements in research speed and software quality while preserving human review and reproducibility.

What you’ll bring

Experience with ML engineering, agent tooling, or research automation. Bring the ability to design careful experiments, build working systems, and distinguish measured progress from misleading evaluation results.

Compensation

We offer a base salary of $200,000–$800,000 and a meaningful equity grant, depending on experience and background, along with competitive benefits.