NewCo – Applied ML / Data Scientist
Data scientist / applied machine learning researcher with expertise in modeling, deep learning, running large scale experiments, preferably having deployed production systems at large companies. AutoML experience is a plus.
Ideal candidate would also have both applied ML and data science experience and some academic background (which suggests experimental rigor).
Open source contributions to projects like AllenNLP, Prophet, Neural Prophet, Hugging Face / BERT, Deep Graph Library, BoTorch are good indications + people who worked at deep learning frameworks directly (TensorFlow, PyTorch, Jax).
Kaggle grand masters with also real world expertise can be an interesting pool.
NewCo is reinventing how people without deep knowledge of machine learning and data science use them for their business use cases.
Our platform builds on two open source deep learning and infra projects we authored (Ludwig and Horovod more than 18k GitHub stars combined) and adds ease of use nad automation that enables expert data scientists to have full control and data analysts to obtain predictions and run analyses in a simple and intuitive way.
We want to make the latest developments in AI accessible also to those who wouldn’t know how to write PyTorch or TensorFlow code themselves, and empower them with a next gen tool.
About the Role
The candidate will be the responsible for the machine learning models we will put in production, making them accurate and efficient, benchmark on a large scale and improve them and creating reusable and automated workflows for data cleaning and processing pipelines, hyperparameter optimization loops and evaluation, model comparisons and reporting.
The candidate will also be responsible for the development of new AutoML features, hyperparameter optimization, new model compression techniques and both supervised and unsupervised model training and tuning, while keeping up with the most recent developments in research.
Finally the candidate will be responsible for developing an entire new breed of multi-modal multi-task models that can apply across many use cases.
What You’ll Do
- You’ll train models on a wide variety of datasets on multiple modalities (taburlar, text, images, audio) and you’ll benchmark them thoroughly in a reproducible way.
- You’ll interact constantly with machine learning and AI technologies from our deep learning and data stacks (TensorFlow, PyTorch, Horovod, Transformers, Ludwig, Ray)
- You’ll keep up with the latest developments in research and implement new models in a modular and reausable way
- You’ll automate processes such as hyperparameter search and AutoML in general
What You’ll Bring
- Experience data science and applied machine learning expertise
- Deep learning and machine learning knowledge
- You are able to deploy machine learning models for use in real world an you can own the the entire data science / ML lifecycle from identification of the problem to monitoring of the deployed solution
- Model design and independent decision making on experimentation and benchmarking
- Experience using, contributing to, and building indestry standard machine learning and deep learning frameworks like TensorFlow, PyTorch or Ludwig
- (Plus) experience with AutoML, hyperparameter optimization and neural architecture search
- (Plus) experience with ML research topics like multi-modal learning, multi-task learning, graph representation learning and self-supervision
- (Plus) experience with model productionization
About Our Process
Our process aims at making sure the candidates can show their strenghts through a number of interviews and exercises to work on together.
- A code screening interview
- A design interview
- A technical communication interview
- A deep-dive pair-coding exercise in building a model together
- Someone with solid industry experience as an IC, ideally with academic experience (papers in conferences like NeurIPS, ICML, KDD, ACL, EMNLP, CVPR, ICCV, UAI, IJCAI, AAAI), that reached their growth / learning ceiling in their current company and looks for new stimuli.
- Someone who is excited about machine learning and AI applied to real-world problems in a customer oriented way.
- Someone who wants to work with a team of veterans, well known in the ML/AI industry.
- Someone who is interested in contributing to a company with open source foundations (maybe you are open soruce contributors / maintainers yourself)
- Someone who is excited about the mission of creating tools to bring powerful technologies to the rest of us
- Someone who can work independently and is self motivated by curiosity and likes an intellectual challenge, but is a team player and puts the success of the team above their own
- Someone who is ok with a hybrid remote + physical workplace
- Someone who sees diversity as a bringer of value rather than a cosmetic badge of honor
Company strong points
- The company has a stellar founding team, made of people with multiple startup exits, creators of well known and adopted open source projects and people from prestigious institutions and companies.
- The company is built on top of years of work on open source technologies.
- The company got started recently, so opportunity to get in early and steer decision making in terms of both culture and technology.
- The company closed a seed round with top tier investors (Greylock) and a new found created by industry experts (The Facory) and has a stellar set of advisors (directors, industry veterans and people with multiple exits).
- The company already has a convincing demo of the tech (no painted dors) and is already in conversation with multiple potential customers.