Staff Software Engineer - Experimentation

About the Role

We are looking for a Staff Software Engineer to own how Headway runs experiments. "Ship small, learn fast" is a core principle, and experimentation is how we achieve it: eleven teams have run hundreds of experiments deciding how patients find a therapist, how quickly new providers get up and running, whether our AI products are working, and how we collect clinical outcomes data that proves care is effective. You will own the platform all of this runs on.

This is a hands-on technical leadership role with unusually wide reach. You will design and build the core primitives, including how users are assigned to experiments, how we record what someone actually saw, and how we recognize the same person across devices and between browsing anonymously and signing in. You will decide what our long-term experimentation and feature flagging stack should be; we run LaunchDarkly and Statsig today, and that architecture is yours to evaluate and consolidate. You will also make AI a real part of the workflow, from validating an experiment's setup before it starts to catching a broken experiment automatically instead of discovering it in a Slack thread a week later.

You will also evolve our experimentation capabilities to expand into market-based testing. Headway is a multi-sided marketplace, so we frequently need to change something for therapists and measure what happens to patients, which changes both how a test must be designed and how it can be analyzed. We need to test whole geographic markets, not just individual users. We need to hold groups out for long stretches to understand real long-term impact. These capabilities are well understood in the industry and we have not built them yet, and they are the ones our business needs next.

The job is a platform and a program in equal measure. Half of it is the technical foundation. The other half is raising how well the whole company experiments: the defaults, the review, the documentation, and the education that let a team run a trustworthy experiment without needing you in the room. You will gather needs from teams who all want something different, sequence them into a roadmap, and say no clearly, which means acting as part technical product manager for your own platform. You will sit in the Patient org, where the highest volume of experiments happens, and serve the whole company.

Trustworthy results are shared ownership. You will own assignment, exposure, and the experiment data pipelines, which are the parts of the chain that are engineering problems. You will work closely with Data Science, who own statistical methodology, and with the teams who own our event tracking and metric definitions, because a number is only as trustworthy as the weakest link in that chain.

Responsibilities

• Design and build core experimentation primitives, including user assignment, exposure tracking, and cross-device identity resolution.
• Evaluate and consolidate our long-term experimentation and feature flagging stack (currently LaunchDarkly and Statsig).
• Integrate AI into the experimentation workflow, such as validating experiment setups and detecting broken experiments automatically.
• Expand experimentation capabilities to support market-based testing for our multi-sided marketplace.
• Develop and maintain data pipelines for experiment analysis and reporting.
• Act as a technical product manager for the experimentation platform, gathering needs, sequencing roadmaps, and making prioritization decisions.
• Raise the experimentation bar across the company through defaults, templates, reviews, and education.
• Collaborate with Data Science and other teams to ensure trustworthy and accurate experiment results.
• Drive technical direction through design documents, reviews, and planning cycles, while mentoring senior engineers.
• Flex across the stack, touching backend services, front-end rendering, CDN behavior, and the data warehouse.

Requirements

• Staff-level engineering experience with a track record of impact measured by what other teams can achieve.
• Experience owning an experimentation or feature flagging stack end-to-end, including accountability for result accuracy.
• Hands-on experience building critical components like persistent assignment, accurate exposure tracking, cross-session identity resolution, or diagnostics for catching broken experiments.
• Fluency in statistics to partner effectively with data scientists, distinguishing between weak and invalid results.
• Experience debugging complex measurement issues, such as uneven traffic splits, cross-arm contamination, or identity resolution failures.
• Proven ability to build a program—not just a platform—that measurably improves how an organization experiments through defaults, templates, reviews, and teaching.
• Experience making build vs. buy decisions for platforms and understanding vendor risks and trade-offs.
• Success in driving migrations on teams that didn’t ask for them, by making the new path easier than the old one.
• Clear opinions on where AI can and cannot be trusted without human oversight, and experience using AI as leverage in your work.
• Experience driving technical direction beyond your own team through design documents, reviews, and planning cycles.
• Willingness to work across the stack, including backend services, front-end rendering, CDN behavior, and data warehouses.

Experience with multi-sided marketplaces, growth or search and ranking systems, or healthcare is a plus, though not required.

About Engineering at Headway

Building a new mental healthcare system at Headway is only possible because of the scale and leverage that software can provide. The engineering team at Headway is a small but mighty team using technology to build that future (and have a fun time while doing it!).

Principles That Guide Us

• Focus on the mission: We view software as a means of effecting change in the world, not as an end unto itself. We write software to empower our patients and therapists to better solve the problems they're facing.
• Ship small, learn fast: We are building new and novel products and believe that we learn what our users need by quickly shipping and iteration.
• Everything is a product: Whether it's our patient search experience or a developer productivity improvement, we treat everything we build at Headway as a product with end-users in mind.
• Optimize for trust: We believe that engineers should be continually learning. To learn effectively and to be productive, engineers must feel safe asking questions and discussing mistakes.

Tools We Use

• Languages: Python 3, TypeScript
• Libraries & Frameworks: FastAPI, SQLAlchemy, React, Remix
• Experimentation & Analytics: LaunchDarkly, Statsig, dbt, Snowflake
• Datastores: Postgres, Redis
• Infrastructure: AWS (Fargate, ECS, S3, and more), Cloudflare, Spark and Kafka
• Monitoring: Datadog, PagerDuty, Sentry
• Version Control: Github

Compensation and Benefits

The expected base pay range for this position is $265,000-$331,000, based on a variety of factors including qualifications, experience, and geographic location. In addition to base salary, this role may be eligible for an equity grant, depending on the position and level.

Benefits

• Equity compensation
• Medical, Dental, and Vision coverage
• HSA / FSA
• 401K
• Work-from-Home Stipend
• Therapy Reimbursement
• 16-week parental leave for eligible employees
• Carrot Fertility annual reimbursement and membership
• 13 paid holidays each year as well as a Holiday Break during the week between December 25th and December 31st
• Flexible PTO
• Employee Assistance Program (EAP)
• Training and professional development