Эта вакансия старше 7 дней и может быть неактуальной.
Чтобы не пропустить новые вакансии и откликаться в числе первых, подпишитесь на уведомления в Telegram
1 month ago
Lead ML Engineer
Описание вакансии
Principal ML Engineer
BACKGROUND
Doublepoint creates cutting-edge interaction technology for next generation user interfaces on smartwatches, TV’s and AR. Our smartwatch algorithms detect subtle, intuitive hand gestures using only built-in sensors like the IMU and PPG — something that hasn’t previously been possible.
To expand the number of gestures we support and push our current set closer to perfection, we’re growing our Algorithm Team. We’re looking for a Principal ML Engineer with deep time-series expertise and strong technical leadership to help own and develop our core product: the gesture classification algorithm.
RESPONSIBILITIES
As Principal ML Engineer, your primary responsibility is technical. This means owning the design, experimentation, performance and deployability of our gesture classification model as a whole which may be at varying levels of technological maturity. This includes:
* Assessing and understanding the feasibility of classification tasks with current signals
* Defining meaningful test sets and metrics, and maintaining both a general and detailed understanding of model performance
* Defining, requesting, and curating the right training data or hardware modifications needed to improve model quality
* Guiding ML Ops engineers on building the required scalable deployment and evaluation infrastructure
* Leading efforts to deploy models securely on embedded compute across a range of sensor hardware
As a technical leader on the team, you’ll also:
* Provide structure, processes, and technical direction for the Algorithm Team when needed
* Help coordinate sprint planning and prioritize work aligned with company-wide quarterly goals
* Mentor junior developers and coordinate the work of relevant freelancers
In addition, you’ll collaborate cross-functionally with:
* The data collection team to define effective dataset collection strategies
* The user research team to help define new gesture types
* The hardware team to inform future revisions of our hardware sensing stack
What’s not your responsibility:
* Implementing hardware
* Pitching to clients
* Leading or implementing ML Ops infrastructure
* Owning the full tech roadmap
REQUIREMENTS
* 5 or more years of machine learning engineering experience in production, ideally in a real-time, time-series and user-centred environments.
* Deep scientific understanding and proven track record with Machine Learning including supervised and unsupervised learning, deep learning, data augmentation, classification, and embedded deployment.
* Strong software engineering skills in Python, Pytorch, TensorFlow Lite, and familiar with scalable ML pipeline tools like Kubeflow, MLflow, Optuna, Hydra and CICD workflows.
* Experience with high accuracy models where high recall and low false positives is critical.
* Experience with signal and model performance variability due to different sensor manufacturers, user anatomies, or usage environments.
Our ideal candidate would also possess
* Familiarity with human computer interaction, and both online and offline testing
* Familiarity with hardware sensors and embedded systems such as IMUs, PPGs, ARM M4 processor architectures, and Tensorflow Lite
* Defining hardware and datasets for models
* Experience with basic signal processing, bayesian statistics.
QUALITIES
* Cares deeply about both model performance and the end user experience
* High agency including taking initiative, owning challenges and outcomes
* Adaptability to work on both high-level strategy and detailed implementation
* Clear communication to articulate complex ideas to technical and non-technical stakeholders.
HIRING PROCESS
BACKGROUND
Doublepoint creates cutting-edge interaction technology for next generation user interfaces on smartwatches, TV’s and AR. Our smartwatch algorithms detect subtle, intuitive hand gestures using only built-in sensors like the IMU and PPG — something that hasn’t previously been possible.
To expand the number of gestures we support and push our current set closer to perfection, we’re growing our Algorithm Team. We’re looking for a Principal ML Engineer with deep time-series expertise and strong technical leadership to help own and develop our core product: the gesture classification algorithm.
RESPONSIBILITIES
As Principal ML Engineer, your primary responsibility is technical. This means owning the design, experimentation, performance and deployability of our gesture classification model as a whole which may be at varying levels of technological maturity. This includes:
* Assessing and understanding the feasibility of classification tasks with current signals
* Defining meaningful test sets and metrics, and maintaining both a general and detailed understanding of model performance
* Defining, requesting, and curating the right training data or hardware modifications needed to improve model quality
* Guiding ML Ops engineers on building the required scalable deployment and evaluation infrastructure
* Leading efforts to deploy models securely on embedded compute across a range of sensor hardware
As a technical leader on the team, you’ll also:
* Provide structure, processes, and technical direction for the Algorithm Team when needed
* Help coordinate sprint planning and prioritize work aligned with company-wide quarterly goals
* Mentor junior developers and coordinate the work of relevant freelancers
In addition, you’ll collaborate cross-functionally with:
* The data collection team to define effective dataset collection strategies
* The user research team to help define new gesture types
* The hardware team to inform future revisions of our hardware sensing stack
What’s not your responsibility:
* Implementing hardware
* Pitching to clients
* Leading or implementing ML Ops infrastructure
* Owning the full tech roadmap
REQUIREMENTS
* 5 or more years of machine learning engineering experience in production, ideally in a real-time, time-series and user-centred environments.
* Deep scientific understanding and proven track record with Machine Learning including supervised and unsupervised learning, deep learning, data augmentation, classification, and embedded deployment.
* Strong software engineering skills in Python, Pytorch, TensorFlow Lite, and familiar with scalable ML pipeline tools like Kubeflow, MLflow, Optuna, Hydra and CICD workflows.
* Experience with high accuracy models where high recall and low false positives is critical.
* Experience with signal and model performance variability due to different sensor manufacturers, user anatomies, or usage environments.
Our ideal candidate would also possess
* Familiarity with human computer interaction, and both online and offline testing
* Familiarity with hardware sensors and embedded systems such as IMUs, PPGs, ARM M4 processor architectures, and Tensorflow Lite
* Defining hardware and datasets for models
* Experience with basic signal processing, bayesian statistics.
QUALITIES
* Cares deeply about both model performance and the end user experience
* High agency including taking initiative, owning challenges and outcomes
* Adaptability to work on both high-level strategy and detailed implementation
* Clear communication to articulate complex ideas to technical and non-technical stakeholders.
HIRING PROCESS
Источник - Startup Jobs in Finland