About the role

Machine Learning Engineer at Winnow developing AI solutions for food waste reduction. Collaborate with cross-functional teams and leverage cutting-edge technologies in food recognition.

Responsibilities

Designs tests and experiments to build machine-learning and AI models for detecting and recognising food and non-food items from images and videos and optionally texts, not excluding other contextual data like date, time, season, geography etc to improve the performance of the models.
Maintains a high level of data quality via directing and guiding the annotation team and third-parties in data annotation and labelling, including leveraging modern AI-assisted annotation approaches.
Manages the data efficiently for model training and evaluation.
Applies state-of-the-art model architectures and techniques (including deep learning, transformer-based and multimodal models) to improve the models.
Prepares reports and presents results to summarise main findings and conclusions
Presents results of scientific research to stakeholders and may contribute to external publications or technical articles.
Writes software and algorithms for training and running the models within Winnow applications.
Collaborates with other teams at Winnow in deploying the models to Winnow's applications to embedded systems like NVIDIA Jetson devices and to the cloud platforms such as AWS and GCP.
Explores and integrates Large Language Models (LLMs) and Vision-Language Models (VLMs) for tasks such as annotation support, detection, multimodal reasoning, and building AI-powered services and agents.

Minimum Master's degree in Machine Learning, Computer Science, Mathematics, Statistics or equivalent with preferably some commercial experience.
Experience in developing and deploying ML and AI models end-to-end gained in at least one corporate environment.
Strong knowledge of one or more of the following areas: Object Detection, Image Classification, Action Recognition, Image Generation, or Semantic Segmentation. Experience in Bayesian methods, Reinforcement Learning, Signal Processing or NLP is a plus.
Experience with modern deep learning approaches, including transformer-based models and multimodal systems (LLMs/VLMs).
Expertise in working with TensorFlow or PyTorch.
Experience adapting modern model architectures to solve real-world problems.
Familiarity with Linux and AWS.
Experience with model inference and optimisation tools (e.g. ONNX, TensorRT or similar). Experience with modern LLM serving frameworks (e.g., vLLM, Ollama) is a plus.
Good programming skills using Python.
Comfortable working independently, prototyping solutions, and bringing them to production.
Teamwork, knowledge transfer, process documentation.
Passionate about Machine Learning and Artificial Intelligence.

Competitive base salary
Meal tickets - 40 RON per working day
2 Wellness hours per month plus a 274 RON gross monthly wellness allowance or the option to swap the wellness allowance for a 7Card subscription
25 days of paid vacation time in addition to national holidays, plus the option to buy a further 5 days annual leave
Company part-funded private health insurance and eye care allowance
Life insurance (3 times base salary)
Company stock options package
Eligible for discretionary annual bonus
Employee Assistance Programme - 24/7 helpline for your wellbeing
Learning and development allowance of 1,730 RON annually
Hybrid way of working - we’re all in the office on Wednesdays and Thursdays
Company provided breakfast & snacks on office days
Early Finish Fridays - log off at 3 PM on a Friday if you have completed your tasks by then
Our own office space with a great working environment
You will love what you do – waking up every day solving one of the biggest social problems of our generation - food waste
Committed team members with broad experience who share a common passion to build a world class business