Machine Learning Research Engineer
The company
We are a rapidly growing startup developing solutions that utilise Virtual Agents to handle manual customer and marketplace operations tasks. Virtual agents blend deterministic Python code, LLM reasoning and agentic AI capabilities to undertake this work, along with a fallback to human experts.
Our unique approach combines the strengths of human expertise (high accuracy and nuanced decision-making) with the advantages of AI automation (speed and cost efficiency). This cutting-edge technology helps businesses solve real-world challenges in trust & safety, and beyond, without the need for complex technical integration. We believe in an online world free from harm, where we can trust AI to make safe and fair decisions.
We have raised about $25M in VC funding from top-tier funds, including Creandum and Plural, and operate at significant scale - analysing millions of daily images and videos from 10+ Enterprise customers. But we are just at the beginning of our journey - and we are very excited about our plans for growth over the coming year and beyond!
The role
We are now looking for a Machine Learning Research Engineer to help build and deliver a platform that can automatically create Virtual Agents for operational processes currently undertaken manually using browser-based UIs. An important part of our offering is the ability to interact with a customer’s existing tooling, and we have found encoding repetitive interactions as Python code to be a powerful strategy. We envision the need for an “Agent Factory” that can compile the deterministic portions of workflows into reliable, testable code. This “Agent Factory” will learn from captured demonstrations of workflows and transform our customers' manual processes into automated solutions that combine the speed and reliability of code with the power of AI, where reasoning is needed.
Your mission will be to create capabilities that automate the creation and management of Virtual Agents. You will use your knowledge of Agentic approaches for code generation and software engineering best practices to design and develop these capabilities - leveraging state-of-the-art LLM coding frameworks and endowing them with the tools and guardrails needed to reliably build and test automated workflows. You will work with customer-facing technical teams to configure and deploy the Virtual Agents for customer work.
You will push the boundaries of Agentic automation, leveraging the best-in-class capabilities where it’s appropriate and developing in-house when we need to.
You will work with platform and software engineers to turn these capabilities into robust operational systems that can scale to deliver Virtual Agents to most operational processes currently undertaken by humans.
As part of this role, you will:
* Design and build capabilities that create Virtual Agents - an “Agent Factory”
* Robustly evaluate the agent creation process, allowing a systematic improvement in capabilities through experimentation against benchmarks
* Implement code generation capabilities, guardrails and evaluations, specialised for encoding workflows based on customer documentation and input, such as captured browser demonstrations
* Drive the uptake of these capabilities with our customer-facing technical teams
* Utilise best-in-class capabilities to deliver these capabilities
* Research, invent and create novel capabilities where gaps in industry require it
* Participate in our machine learning community to influence how we implement machine learning and computer vision technologies, shaping Unitary's future.
* Contribute full-stack development, including software engineering, DevOps, and MLOps, along with light task and project management to ensure your capabilities deliver maximum value early.
We would love to hear from you if you:
* Know how to create systems for Agentic development, including mechanisms to guide and enhance state-of-the-art LLMs
* Have expert knowledge of the capabilities of Agentic AI and Generative AI
* Can assess where best-in-class industry capabilities can help us undertake operational workflows
* Know how to invent novel capabilities based on rapid research iterations
* Can work with other engineers to understand and solve challenges
* Have strong Python and engineering skills
* Can demonstrate problem-solving and project management skills to analyse workflows and design automated solutions
* Thrive in a collaborative environment where group output and team achievements weigh heavily than individual input
* Can travel to our company-wide off-sites three times per year
It would be even better, but not essential, if you have:
* Experience working in a fully remote, international team
* Experience with Temporal or similar workflow orchestration platforms
* Previous startup experience
* Experience with MLOps practices and tools, and monitoring machine learning systems in production
* Knowledge of browser-based automation methods, such as playwright
* Knowledge of CI/CD practices and tools such as GitLab CI, Argo CD
* Proficiency with SQL and NoSQL databases
* Worked with Kubernetes and infrastructure as code (IaC) tools such as Terraform
This role will report to the Chief Scientist and can be placed anywhere within 3 hours of the UK time zone.
The team
Unitary is a remote-first team of c. 20 people spread across Europe and North America who are fiercely passionate about making the internet a safer place, and deeply motivated to become a force for good. We have an ambition to create a company filled with happy, kind and collaborative people who achieve extraordinary things together. Our culture is built around the power of trust, transparency and self-leadership.
Working at Unitary
We are committed to creating a positive and inclusive culture built on genuine interest in each other's well-being. We offer progressive and market-leading benefits, including:
* Flexible hours and location
* Competitive salary and equity package
* Occupational pension
* Generous paid parental leave
* Generous paid sick leave
* Annual budget for your professional development and growth
* Annual budget for your individual health and wellness
* Three team off-sites to London or other exciting destinations in Europe