AI Engineer
- $227,000–$284,000 per year
- Hybrid
- Full Time
About the role
Superhuman offers a dynamic hybrid working model for this role, requiring 3 days per week in our SF Hub. This flexible approach gives team members the best of both worlds: plenty of focus time along with in-person collaboration that helps foster trust, innovation, and a strong team culture.
About Superhuman
Grammarly is now part of Superhuman, the AI productivity platform on a mission to unlock the superhuman potential in everyone. The Superhuman suite of apps and agents brings AI wherever people work, integrating with over 1 million applications and websites. The company’s products include Grammarly’s writing assistance, Superhuman Docs's collaborative workspaces, Mail’s inbox management, and Go, the proactive AI assistant that understands context and delivers help automatically. Founded in 2009, Superhuman empowers over 40 million people, 50,000 organizations, and 3,000 educational institutions worldwide to eliminate busywork and focus on what matters. Learn more at superhuman.com and about our values here.
The Opportunity
We’re looking for a Senior AI Engineer who takes real ownership and does well in a fast-moving, ambitious, and problem-oriented environment. You’ll work hands-on with improving our LLMs, and using them to solve for high-impact product opportunities. You’ll own part of the direction of our AI work: framing the right problems, finding their most effective solution, and helping shape the roadmap.
This role is a bit of a hybrid of machine learning/engineering with product. It would suit someone already in an AI Engineering role, as well as software engineers or data scientists that have gone deep into shaping models (themselves) to propel them forward.
What you’ll do
Lead entire AI/ML initiatives from problem definition through delivery.
Influence technical direction and roadmap decisions by framing problems clearly and backing choices with evidence.
Partner with product, engineering, and design to deliver high-impact models.
Strengthen our ML practices and infrastructure, including evals, training pipelines, and model serving.
Mentor teammates and share knowledge to raise the bar across the team.
By 30 days, you will have:
Upgraded an existing feature to be meaningfully smarter and/or faster, and shared what you learned with the team.
Reviewed our evaluation approach and improved or built evals for our core features.
Found the highest-leverage opportunities in our AI stack and started shaping priorities with the team.
By 90 days, you will have:
Led the delivery of AI features that address valuable customer use cases.
Fine-tuned models for our specific needs and set repeatable practices others can build on.
Developed new processes to speed up how we utilize new AI models.
Qualifications
A track record of owning and shipping production AI/ML systems end-to-end, with the independence and judgment expected of a senior engineer.
A blend of modern LLM expertise and structured problem-solving. You can work backward from a user problem, frame the right technical question, and define how to evaluate success. That structured problem-solving might come from product engineering, technical consulting, or comparable experience.
Deep hands-on experience in LLM post-training and/or inference optimization, with evidence of exceptional results in a specific problem area. For example:
Building a model that outperforms strong alternatives on a specialized task through thoughtful dataset development, supervised fine-tuning, preference tuning, or reinforcement learning.
Making LLMs substantially faster or more efficient through tailored quantization, speculative decoding, or other inference techniques.The ability to go beyond established recipes: diagnosing failure modes, designing experiments, and adapting methods to the problem at hand.Strong evaluation judgment. You design evals grounded in user needs, distinguish meaningful improvements from benchmark gains, and establish practices that catch regressions before users do.
Experience serving and optimizing models at scale, with a practical understanding of the trade-offs between latency, throughput, uptime, and cost.
Strong production Python and fluency with data, including building and cleaning datasets, analyzing results in Python or SQL, and working comfortably with modern training and serving stacks.
Experience influencing technical direction and helping other engineers succeed through clear reasoning, collaboration, and mentoring.
Nice to have
Experience spanning both LLM post-training and inference optimization.
Successful application of preference tuning or reinforcement learning to improve real-world model performance.
Speech or long-context experience, especially with noisy, lengthy transcripts.
Experience establishing ML infrastructure, evaluation practices, or model development workflows at a startup or growth-stage company.
How you work
Ownership: You take ambiguous problems and drive them to clear outcomes without waiting for perfect specs.
High standards: You care about quality and detail, you’re tenacious, and you raise the bar for everyone around you.
Pragmatic and results-driven: You pick the right hills to climb and find the fastest way up.
A generalist who dives deep: You might work on data one day and training methods or infrastructure the next.
Clear communicator: You explain technical results and tradeoffs clearly to both technical and non-technical partners.
Multiplier: You make the people around you better through mentorship, reviews, and knowledge-sharing.
Compensation and Benefits
Superhuman offers all team members competitive pay along with a benefits package encompassing the following and more:
Excellent health care (including a wide range of medical, dental, vision, mental health, and fertility benefits)
Disability and life insurance options
401(k) and RRSP matching
Paid parental leave
20 days of paid time off per year, 12 days of paid holidays per year, two floating holidays per year, and flexible sick time
Generous stipends (including those for caregiving, pet care, wellness, your home office, and more)
Annual professional development budget and opportunities
Superhuman takes a market-based approach to compensation, so base pay may vary by location.
For more information about our compensation zones and locations where we currently support employment, please refer to this page. If a location of interest is not listed, please speak with a recruiter for additional information.
Base pay may vary considerably depending on job-related knowledge, skills, and experience. The expected salary ranges for this position are outlined by compensation zone and may be modified in the future.
We encourage you to apply
At Superhuman, we value our differences, and we encourage all to apply—especially those whose identities are traditionally underrepresented in tech organizations. We do not discriminate on the basis of race, religion, color, gender expression or identity, sexual orientation, ancestry, national origin, citizenship, age, marital status, veteran status, disability status, political belief, or any other characteristic protected by law. Superhuman is an equal opportunity employer and a participant in the US federal E-Verify program (US). We also abide by the Employment Equity Act (Canada).