Machine Learning Scientist
Eli Health · Montreal (in-person)
Extracted from the posting
Applied Machine Learning Scientist role on the ML team, focused on open-ended scientific problems with noisy, imperfect real-world datasets rather than ML infrastructure. Half the time is spent on analysis and modelling (classical ML, signal processing, statistics, deep learning where appropriate), half on production performance, error analysis and root-cause investigation. Work is in Python, with reproducible code not limited to notebooks; the ML lead handles productionization.
About us
Eli Health is making continuous hormone monitoring possible so users can support their daily and long-term health. No more waiting days to track key biomarkers—get results you can use within minutes. Eli’s flagship product, Hormometer®, is the first instant hormone monitoring platform to deliver results from saliva to mobile app—anytime, anywhere.
Developed over six years of R&D, over 2,000 product iterations, and backed by a dozen patent-pending innovations, Eli’s award-winning platform turns hormones into measurable signals you can track and improve. Just as the thermometer and glucometer have transformed health for millions, Eli’s platform is poised to be the next major evolution in tracking changes in stress, endurance, sleep, and more.
About the role
Eli is looking for an on-site Applied Machine Learning Scientist to join our ML team and help solve challenging real-world problems involving imperfect, noisy, and complex datasets.
This role is primarily about scientific problem solving and applied machine learning, not ML infrastructure. You will work closely with the ML lead, who will support integration and productionization of your work
Where you'll spend the first half of your time
You’ll work on open-ended scientific and machine-learning problems, turning imperfect real-world data into robust, reproducible analyses and models. The results of these analyses will have to be communicated effectively to the broader team. The focus is on understanding the problem deeply, choosing the right methods, and validating that improvements are real and generalizable.
- Explore datasets and understand the underlying data-generating processes.
- Develop, evaluate, and improve machine learning, signal processing, and statistical models.
- Perform feature engineering, model selection, validation, and error analysis.
- Identify issues such as confounding, data leakage, measurement variability, and distribution shift.
- Design experiments and analyses to resolve uncertainty and guide modelling decisions.
- Investigate new modelling approaches, including classical ML and deep learning (when appropriate).
- Produce clear, reproducible Python code that isn’t limited to notebooks, and communicate findings to technical and non-technical stakeholders.
Where you'll spend the other half
You’ll stay closely connected to how models behave in the real world. You’ll investigate production performance, diagnose failures and unexpected behavior, and use those observations to drive new analyses, experiments, and model improvements.
- Develop analyses and models with the expectation that it will be deployed into production.
- Investigate model performance and failures using real-world production data.
- Perform ongoing error analysis, diagnostics, and root-cause investigations.
- Identify distribution shifts, edge cases, systematic biases, and degradation in model performance.
- Translate production observations into experiments, improvements, or data-collection strategies.
What we are looking for
Someone with a bachelor's degree (Master’s or PhD preferred) in Engineering, Computer Science, Data Science, Mathematics, or a related field who possesses a minimum of 5 years of professional experience excluding internships. Additional capabilities include:
- Strong foundation in machine learning fundamentals, statistics, and experimental reasoning.
- Strong Python skills and experience with common ML/data science libraries.
- Ability to work independently on ambiguous problems and determine what questions need to be answered.
- Good understanding of model validation, uncertainty, bias/variance, and generalization.
- Ability to distinguish between improvements that are statistically or scientifically meaningful and those that simply improve a metric.
Above all, we are looking for someone who is particularly good at answering the following question:
<code>Given the data we have and the problem we are trying to solve, what can we conclude with confidence, what remains uncertain, and what should we do next?</code>
Why you’ll love working at Eli
- You’ll work with a group of talented and mission-driven people eager to improve lifelong health at scale.
- You’ll be part of the core team developing and commercializing the first product that monitors hormonal data daily and over a lifetime.
- You’ll join the early-stage startup phase and have a wide-reaching impact in a constantly evolving, fast-paced environment.
- You’ll be part of a small (<25 people), high-performing, and diverse team where everything you do results in tangible impact and shapes the company’s trajectory.
- You’ll be in an environment where people drive their own work, think creatively about open-ended problems, and solve them proactively.
- You’ll get health insurance (medical, dental, vision, and more) to ensure you and your family stay physically and mentally at your best.
- You’ll have flexibility over your schedule and vacations. We seek to hire great people, then give them the autonomy and space they need to achieve their goals.
- Although our office and R&D facilities are in Montreal, we have a distributed team. We prioritize asynchronous workflows and minimize meetings to focus on the work itself.
- You and your +1 will have unlimited free access to the Bota Bota spa in Montreal to recharge.
How to apply
Sounds like you? Please apply using this link:https://jobs.ashbyhq.com/eli. Please note that we will not review applications made through other platforms. We look forward to learning about you!
