Who you are

• 5+ years of experience optimizing and productionizing modern ML models, especially ones that run in a real-world product environment (bonus if you’ve worked closely with transformer models)
• You have deep experience fine-tuning open-source LLMs and going beyond simple LoRA fine-tuning
• You have production experience with a modern coding language like Python
• You're passionate about on-device performance and excited to push the boundaries of what's possible in a browser
• You have experience independently running critical projects, shipping ML features, and leading initiatives with minimal guidance
• You’re pragmatic, motivated by nebulous problems, and excited to work in a startup environment with quick product validation cycles
• We’re primarily focused on hiring in North American time zones and require that folks have 4+ hours of overlap time with team members in Eastern Time Zone

What the job involves

• As a Staff ML Engineer you’ll be working alongside ML engineers, product engineers, designers, and our cofounder and CTO, Hursh Agrawal, to build the next LLM-powered interface for the internet
• You’ll work with designers and product engineers to define the product, fine-tune LLMs and smaller transformers like BERT, and iterate on our how we host models both in the cloud and on-device to improve latency and resource usage
• Fine-tune, distill, and optimize LLMs to improve performance, reduce latency, and enhance efficiency for on-device and cloud-based inference
• Improve our on-device model architecture, leveraging frameworks like MLX, ONNX, and TFLite to ensure models run efficiently across different devices
• Set product direction via your expertise in ML and how it can help power Dia and increase it’s capabilities
• Keep up with the latest advancements in AI and apply new models and techniques to improve Dia where applicable
• Build evaluation pipelines to track model performance, accuracy, and real-world effectiveness over time. You'll also collaborate with product ops teams to build and improve datasets that accurately match product needs
• Iterate on the product experience by integrating and improving the models used in Dia, working alongside product engineers and designers
• After 1 month you will..
• Onboard to the team and codebase with your onboarding buddy
• Attend onboarding presentations about the company, product, codebase, and culture
• Become familiar with the Swift language, the Dia codebase, and how we ship features
• Ship a few bug fixes and small improvements across our product codebase and tooling
• Have trained your first model, either improving an existing flow or enabling an entirely new one and have integrated it yourself into our product codebase
• Have pair programmed with a few people on the engineering team
• Be regularly posting product feedback about the browser in our #dogfooding channel
• After 3 months you will..
• Be familiar with how we prototype and build new features, working with product engineers to brainstorm ways to use models to add intelligence to Dia
• Be familiar with our cloud infrastructure and data pipelines
• Be familiar with how we run inference both on-device and in the cloud
• Be testing new prototypes with existing, on-device models to test performance and viability
• Participate in product brainstorms to think about the future of Dia
• Be trained to interview candidates for roles at the Browser Company
• Be contributing to on-call rotations and jumping into incidents to support the team
• Regularly attend weekly engineering discussions about our architecture, how we do code review, code style, and more
• After 6 months you will..
• Collaborate with our CTO and other ML and infrastructure engineers to shape the product roadmap
• Creatively solve problems with product engineers, using pragmatic solutions ranging from basic heuristics, regressions, ML models, to AI depending on the feature
• Own our on-device model architecture, updating it to try new models, change how we work with LoRA adapters, and optimizing it for performance and quality
• Own our infrastructure to collect training data and fine-tune models for our use-cases
• Have built out mechanisms to assess quality and performance, and be working with product teams to improve the efficacy of our models and heuristics
• Drive projects from conception to production launch independently
• Be mentoring and pair-programming with newer engineers to help them get spun up on the codebase

The application process

• Deadline to Apply: October 30, 2025 at 4:00 AM GMT

Benefits

• Choose-Your-Own-Compensation Plan: When we give you a job offer, we’ll actually give you two different offers — one that is focused on having a higher salary and one focused on higher equity. Candidates can choose either offer (or any numbers in-between) when they accept.
• Best in the Biz Insurance: We provide employees with full coverage medical, vision, and dental insurance and a free One Medical membership to help you keep you and your family healthy. We cover 100% of employee plan costs and up to 90% for dependents.
• Napping During the Workday: Some of us are better in the mornings, others are more productive in the afternoon. People should take breaks when they need to. If you’re on our team, we trust you to get your work done — our policy is: work when it makes sense for you!
• Take Time Off, Please!: We believe that people do their best work when they’ve rested and had valuable time off. So rather than having a maximum number of days off, we actually have a recommended minimum vacation of 20 days a year. Take time off when you need it and also just when you feel like it.
• Extremely Flexible Parental Leave: We are proud to offer an extremely flexible 12 weeks of parental leave — for any parent regardless of gender and for birth, adoption, surrogacy, or fostering. Take the 12 weeks all at once, or split up days to slowly transition back to work — we’re accommodating to everyone’s different situations! Birth mothers also qualify for an additional 6 weeks of medical leave.
• Design Your Dream Home Office: We are a remote first company. About half of our team currently lives outside of New York and across 5 different time zones. So if you need some stuff to make working from home a bit easier, we offer a $1000 stipend for anything you need to get comfortable.
• Cozy Office in Brooklyn, Wooo!: If you do happen to live in the New York area (or you want an excuse to visit) you can come into our beautiful office in Williamsburg. Coming into office is entirely optional, some of us come almost every day and some of us come in once a week. No obligation, we just genuinely like each other!