ABOUT US

We’re building a platform that powers the most advanced AI groups in the world, including Mistral AI, AI21, Meta, and Allen Institute for AI with human feedback data for evaluating and training their models.

Kili was established in 2018 by former machine learning engineer dedicated to delivering the highest quality data in the industry, driving advancements in the field.

Over the past 6 years, Kili has powered nearly every major breakthrough in AI, working for the biggest governments, and the most talented AI teams.

We have built an elite workforce, we have developed our custom multi-modal annotation platform, and implemented sophisticated quality control systems. Our product is a game-changer for frontier ML teams.

ABOUT THE ROLE

As a Data/ML Engineer, you will play a critical role in analyzing data, optimizing data collection pipelines, and ensuring the quality and diversity of data used to train machine learning models. Your work will directly impact the performance and reliability of frontier LLMs, and you will be integral in advancing our objectives across the company. This role requires a deep understanding of data science, a passion for solving complex problems, and the ability to thrive in a collaborative, fast-paced environment.

WHAT WE ARE LOOKING FOR

• Passion for Data Analysis: You enjoy diving into data, identifying patterns, and informing data collection paradigms to enhance model performance.
• Problem-Solving Enthusiast: You are passionate about the challenges we face and are eager to contribute to applied machine learning and data science solutions.
• Team Player: You are excited to join a small, dynamic team and be a key contributor to the company’s success.

KEY RESPONSIBILITIES

• Optimize Data Collection Pipelines: Enhance the efficiency and effectiveness of our data collection processes to improve model performance.
• Implement Quality Control Processes: Develop and refine quality control measures to ensure diverse and unbiased data for robust model training.
• Analyze Existing Data: Investigate current data sets to identify pain points and iterate on machine learning workflows.
• Design Data Collection Workflows: Create workflows tailored to specific model requirements to maximize data quality and relevance.
• Reduce Bias and Ensure Safety: Implement strategies to minimize biases and ensure the safety and reliability of our models.
• Run Experiments: Conduct experiments to refine and improve leading models, making them more insightful, compelling, safe, and aligned with human values.

ABOUT YOU

• You have at least 2-3 years of experience in a consulting firm specializing in data, AI, digital transformation, or a fast-paced company in a customer-facing role.
• You possess a strong foundation in data science and machine learning, with knowledge of Python being mandatory
• You leverage data to drive decision-making and solve complex problems
• You love analyzing data :You'll inform data collection paradigms that improve model performance.
• You care about the problems we’re trying to solve :You're interested in applied machine learning and data science for frontier AI.
• You’re excited about joining a small team : As an early team member you’ll be a key part of advancing objectives across the company.

Join us as a Frontier Data/ML Engineer and be at the forefront of advancing AI capabilities through innovative data analysis and collection strategies. Your contributions will be essential to our mission of producing frontier AI with frontier data.

Kili is an Equal Opportunity Employer. All qualified applicants will receive consideration for employment without regard to race, sex, gender, sexual orientation, age, colour, religion, national origin, protected veteran status or on the basis of disability.

Applying at Kili is also the opportunity to access a broader network. Should we not proceed with a job offer, we would be pleased to refer you to the Talent Club. The talent club was created by Serena and aims at offering talents great opportunities in innovative companies (Dataiku, Malt, Libeo…)