Member of Technical Staff, Research Engineer (Finetuning)
What We're Building
As Inflection embarks on a new stage of growth, we are focusing on collaborating with commercial partners to adapt and fine-tune our cutting-edge models for their unique business requirements. Our accomplishments in developing, aligning, and deploying state-of-the-art models in our high EQ consumer-facing chatbot, Pi, have established a strong foundation for success. Well-funded and equipped with ample H100 resources, we have built a robust infrastructure and efficient processes to support best-in-class finetuning. By joining our team, you'll have the opportunity to contribute your expertise while being part of a dynamic organization that values innovation and collaboration.
About Inflection
Inflection is a small, interdisciplinary AI studio. We have trained several state-of-the-art language models, including Inflection 1 and Inflection 2.5, and built a personal assistant named Pi. As a studio, we are currently focused on finetuning and deploying models for specific use cases for our commercial partners.
We believe that artificial intelligence represents the beginning of an era of exponential change. Our name Inflection embraces this moment of transformation, whilst our status as a public benefit corporation provides us with the legal mandate to prioritize the well-being and happiness of our partners, users, and wider stakeholders above all else.
About the Role
Member of Technical Staff, Research Engineer (Finetuning)
Inflection has trained state-of-the-art models such as Inflection 2.5, but to make these models appropriate for deployment to our enterprise partners, they require finetuning and alignment. Research engineers collect and curate datasets, experiment with novel finetuning techniques, and evaluate the resulting models to ensure they meet our standards for safety, helpfulness, and reliability in enterprise settings.
This is a good role for you if you:
- Have worked on finetuning and evaluating LLMs before, either with in-house models or via API.
- Like working in a fast-paced environment and are comfortable working with fast moving/evolving technical requirements.
- Are comfortable operating on large compute clusters via tools like Slurm.
- Have a strong understanding of modern machine learning techniques (transformer architectures, RLHF, DPO, etc.) and are proficient with PyTorch.
- Knowledge of SQL and data tooling like Snowflake, Dagster, and Airbyte is a bonus, but is not expected.
We do not require a specific university degree or years of experience. Instead we’re excited to see what you’ve been building. Please send us examples of your best work, including but not limited to links to open source contributions, personal projects or a cover letter describing past projects that you are proud of.
Employee Pay Disclosures
At Inflection AI, we aim to attract and retain the best employees and compensate them in a way that appropriately and fairly values their individual contributions to the company. For this role, Inflection AI estimates a starting annual base salary will fall in the range of approximately $175,000 - $325,000 depending on experience. This estimate can vary based on the factors described above, so the actual starting annual base salary may be above or below this range.
How We Work
We value excellence and ownership. Our organizational structure focuses on individual responsibilities rather than management hierarchies. Everyone is expected to lead by doing. We are big believers in the unreasonable effectiveness of highly talented Individual Contributors who are given all the resources, space and ownership to move fast and deliver outstanding results.
Teamwork and generosity are at our core. Our culture celebrates positive challenges, asking questions, learning and actively supporting one another. This mentality of shared respect and purposeful teamwork is key to our success. We equally value all technical and non-technical contributions.
Constructive disagreement is essential. We appreciate when team members challenge assumptions, put forward new ideas, or encourage us to move faster or slower. Openness, honesty and kindness make us great.
Feedback is our ground truth. We have a tight feedback loop between the user experience and our AI creation process. Quantitative and qualitative data drives our priorities. This goes for internal culture too. Everyone has ownership and visibility into key decisions and progress.
Writing creates accountability. Whether on internal communication tools or in team memos, we are strong communicators with a special focus on the written word.
We deeply value time to reset outside of work. We encourage one another to constantly take time to recharge and always focus on maintaining a healthy work-life balance.
Engineering at Inflection
We are a vertically integrated AI studio. This means that our entire technology stack – from large foundational model pre-training to the user interface – is built in-house, with each of the components co-optimized to deliver the best AI experiences. We have built one of the most advanced large language models in the world, based on multiple novel and proprietary innovations.
We believe in scale as the engine of progress in AI, and we are building one of the largest supercomputers in the world to develop and deploy the new generation of AIs.
We wear multiple hats and don’t distinguish between engineering and research. We continuously explore and exploit, creating new and perfecting existing techniques and solutions. User feedback is our North Star.
Our Benefits
We offer generous benefits to ensure a positive, safe, inclusive and inspiring work environment for all Inflectioneers.
- Unlimited paid time off
- Parental leave and flexibility for all parents and caregivers
- Generous medical, dental and vision plans for US employees
- Compliance with country-specific benefits for non-US employees
- Visa sponsorship for new hires
- Avenues for personal growth such as coaching, conference attendance, or specific trainings
Diversity & Inclusion
We are building personal AIs that we hope will serve everyone. We are deeply committed to representing the full extent of the human experience inside our AI Studio. This means that everyone from any walk of life is welcome if you have the right skills. We populate diverse candidate pools for all open roles.