Machine Learning Data Scientist

Key Skills

Java Programming, Python Programming, C++, C Programming, SQL, Cloud computing, Scala Programming, Machine learning techniques, Data science techniques, MATLAB Programming, PyTorch, TensorFlow, R Programming

Job Description

Gordian Biotechnology is a therapeutics company focused on diseases of aging, the major unmet medical need of our generation. Age-related diseases have complex causes that include interactions with the aged environment, and traditional ex vivo screening methods have failed to produce effective treatments.

To address this problem, Gordian’s platform delivers and tests hundreds of therapeutics in individual animals. We thus scale in vivo efficacy testing to cover every therapeutic target at the very beginning of the drug discovery process. This lets us choose targets that really impact disease biology, to develop and commercialize therapies for the world’s deadliest diseases, and eventually for the processes of aging itself.

The Journey:

Our mission is audacious, and the path will be full of both challenges and excitement. Two things characterize the Gordian experience: 1) We work as a team, with ownership in our own roles and trust in each other. 2) We strive for extraordinary outcomes, and in doing so grow our skills and capability.

Team – Relying on each other begins with transparency. We set clear goals, visibly connecting individuals and teams to our company objectives. This empowers each of us to make autonomous decisions about our work, knowing how they will affect the bigger picture. Our communication happens out in the open. We give and receive feedback from a perspective of helping each other grow, share mistakes, and ask for help.

Extraordinary – Every day, we ask ourselves ‘could this process/outcome be even better?’. Knowing our overall mission, we do what we think helps us make the most progress, without asking for permission. We don’t shy away from big challenges or unknown territory, but find a way to excel. Our colleagues are amazing, both at what they do and as people. They inspire us to keep up, to not let them down, and be inspiring in return.

Like any cutting-edge research environment, Gordian doesn't believe in a standard 9 to 5 day. We set ambitious project goals and support each other to meet them, while maintaining the balance individuals need to thrive and achieve excellence. Our team is geared towards helping each other out and maintaining a culture of intellectual and social fun. We like getting things done and keep standing meetings to a minimum. This year, in accordance with our unlimited vacation policy, each member took an average of 2.5 weeks of (offline) vacation, not including major holidays. We unwind with weekly team lunches, and support each others’ projects both experimentally and by ‘pre-mortem’ meetings to discuss possible failure modes and experimental design.

The Details:

Gordian aims to provide everything you need to thrive. Beyond our community and science, you’ll have enough equity to be a true stakeholder in the company, competitive salary, full health/dental/vision/life insurance, 401k with match, whatever vacation you need to be at your peak, remote work flexibility, and access to world-class mentors and advisors to support your professional growth. Our labs are in the Dogpatch, near UCSF Mission Bay.

You have a history of successful execution, both in- and outside of your job description. You want a chance to do your best work, to immerse yourself and excel alongside people who will inspire you and whom you’ll be excited to spend your days with.

You truly want to play a key role in an early-stage startup: A fast-paced environment full of both uncertainty and new challenges, demanding relentless resourcefulness. You have deep expertise, and broad curiosity.

You have experience diving into complex single cell datasets, and quickly finding ways to reap biological understanding in collaboration with other specialists. You are excited to use our unique in vivo data to find better ways to treat complex diseases.

Your mission in this role is to amplify the power of Gordian’s in vivo screening platform, by building and applying machine learning tools for interpretation and out-of-sample predictions of cell states and responses. Gordian has multidimensional data across many cell types, with thousands of biological perturbations, in many species. With your help, this lets us identify not just cell state changes in disease, but the transitions that lead to these outcomes. Our data is unique, and making sense of it will require methodological innovation. You’re excited to use and extend state of the art tools for analyzing cell states from transcriptomic data, in order to draw biological conclusions and guide target discovery. Your work will happen in constant dialogue with both bioinformaticians and biologists; they will bring you biological hypotheses to explore using our data, and you will generate hypotheses which we can test using our screening platform. Output from the tools you build will be used by the whole company, and you’ll design and document them accordingly.

In your first six months you will contribute one major improvement to each of three areas:

• mapping animal perturbation results onto human transcriptomic data to quantify translatability of response,
• identifying likely causal factors in disease-related transcriptomic shifts to prioritize targets to test, and
• identifying transcriptomic features that correspond to specific aspects of pathophysiology to better interpret our single cell data. To do this, you will have access to both public and Gordian proprietary human patient data, as well as data from multiple animal models. We have a broad range of calibration interventions in these models, which will allow you to test hypotheses with well-controlled experimental data. Your one month progress towards these goals will be to establish a framework for testing and adding new data to reference datasets, including automatic annotation by QC metrics, as well as improving Gordian’s current methods for classifying source sample metadata for newly sequenced cells.

Strong candidates will have created one or more software tools for classifying and/or modelling single cell data. You have a strong quantitative background, e.g. from degrees/work in math, engineering, or physics. You’re thoroughly comfortable with python/R, github/gitlab, and PyTorch/TensorFlow/Keras. You combine a mindset of process optimization and reliability with intrepid build-from-scratch creativity. Bonus points for experience with cloud computing environments and/or with workflow languages.