Back to search
Who we are
At SonicJobs we're building the next generation of autonomous web agents based on computer-use technologies. Our agents operate on job application flows, like candidates do. If you've ever dreamed of building AI that actually uses a website, this is your moment.
We're a small, tight-knit team with big ambitions, leveraging bleeding-edge tools to build agents that understand and control web environments end-to-end.
We move fast, avoid unnecessary hierarchy, and aim high. We're not automating job applications - we're redefining how job applications work in order to allow candidates find the job opportunity they deserve.
We are backed by top VC firms and serve more than one hundred corporate clients including several Fortune 500 companies.
About the role
We're looking for an AI engineer to work on the architecture, training, and evolution of our AI agent stack. As Engineer, you'll contribute to fine-tuning the LLMs that will support web-native agents capable of learning, adapting, and thriving in complex job application flows.
This is a demanding, high-impact role. You'll be shipping code, training models and designing experiments. You'll be reporting to the CTO. If you're excited by the idea of coding, reviewing training runs, and designing new RL environments - you'll thrive here.
What you'll do
- Build, train, and deploy LLM-driven agents that control the web via real browser interfaces.
- Fine-tune foundation models for computer-use using techniques like LoRA, PEFT, DPO, RLHF and similar depending on your judgment.
- Design and run reinforcement learning systems to improve agent reasoning and performance.
- Work on the full agent stack: from policy models to interface planning and simulation frameworks.
- Prototype quickly, test aggressively, and constantly push the limits of what agents can do.
- Contribute to the product direction and research priorities.
- Create multi-agent flows with clear trajectories, role assignment, and long/short memory.
What you'll bring
- ML background - PhD or equivalent experience in AI, ML, Computer Science or a strong industry experience.
- Real experience fine-tuning LLMs.
- Deep understanding of Reinforcement Learning (especially applied, not just theoretical).
- Experience building agents in real-world environments (bonus for browser-based or robotic agents).
- You're hands-on: you like writing code, running experiments, and seeing real results.
- You're driven by ownership, not job titles. You're here to build, not manage for the sake of it.
Bonus points for
- Track record of high-impact research published on meaningful or open-source contributions.
- Comfort working in fast-paced, exciting startup environments.
What we offer
- Salary depends on experience.
- Onsite work with a deeply focused, no-BS te
- Real ownership and the chance to shape a frontier space.
- No corporate nonsense - just smart people building ambitious tech.
Sounds like you?
Let's talk. This isn't just a job - it's a chance to lead the future of how job platform technology looks like.
