About Nabla Bio
Antibodies need to possess a multitude of properties to be manufacturable, safe, and efficacious for patients. This is a hard engineering problem with a high a failure rate, because today’s discovery technologies rely on brute-force guess-and-check. At Nabla, we're building a protein design platform that enables us to build antibodies deliberately, with all the properties they need to reach and treat patients, and do so faster than we can today.
Learn more at: https://www.nabla.bio/
About the role
Skills: Biology, Torch/PyTorch, Python, Machine LearningOur mission is to enable pharmaceutical and biotech companies to bring more antibody therapies to patients. Using AI and massively parallel experimentation, we design antibodies that precisely bind the disease target at the right location, while minimizing manufacturability and toxicity risks. We are a well-funded, revenue-generating, bilingual company of wet- and dry-lab scientists, and are founded by AI and protein design experts from Harvard University.
The role
Fueled by partnerships and increasing demand for internal R&D, we will be looking to you to apply and develop ML-guided antibody design technologies in tight collaboration and feedback with our wet-lab. This will include:
- Developing novel strategies and optimizing existing ones to predict antibody function from sequence and structure
- Developing methods to predict and design antibody-antigen interactions
- Develop sequence- and structure-informed representations that enable multi-property antibody engineering
- In collaboration with our wet-lab, designing antibody structures and sequences for functional measurement in frequent design-build-test cycles
Qualifications
- Bachelor’s or master’s degree, with a PhD or equivalent preferred.
- Leading of a multi-month machine learning research project that resulted in a publication, or tool that has been impactful for your previous employer, lab, or other users.
- Strong understanding of statistics and machine learning fundamentals. Practical experience developing deep learning models from scratch, and tuning existing ones.
- Fluency in Python and PyTorch and commonly used higher-level frameworks for model training and hyperparameter tuning.
- Fluency with Unix environments, AWS, and GitHub
- You are problem-focused, and interested in working in a high-intensity, fast-paced environment often driven by deadlines
- You value unblocking colleagues before yourself, and are excited to mentor/train junior colleagues
Technology
Interview Process
A typical interview process looks something like:
- Meeting with Frances or Surge for 30 minutes as an initial screen and to make sure we're on the same page.
- Meeting with Surge or Frances (whomever you didn't meet the first time) to further assess technical and behavioral skills, as well as answer any new questions you had.
- After this step, we'll ask you to sign an NDA. We do this so that we can be open about what we're working on, and so that you have all the information you need to make a decision.
- An on-site interview in Boston with the whole team; runs from ~9AM - 5PM.
- In the morning, we'll present to you for an hour about what we're working on.
- Then you'll give a 30-40 minute job talk to the whole team (usually takes 1-1.5 hours with questions). We'll brief you in advance as to what it should contain.
- Lunch together with the team
- In the afternoon, 1:1s with relevant members of the team.
- Shortly after that, we’ll ask to speak to ~3 references, and we may schedule another 30-45 minute follow-up discussion. This isn't usually another interview, and is more to clarify expectations for the role and answer any other questions you have.
- If we're mutually excited about working together, we'll extend an offer!
We aim to keep our process efficient and not waste either of our time. The whole process can be done in 2-3 weeks.
