Member of Technical Staff, Machine Learning
Bjak · Remote · Poland ·
COMPENSATION
€•• – ••k
Listed on the employer's job board (Ashby)
- ML for mobile apps
- Python + PyTorch/JAX
- Production ML focus
- Small high-talent team
We are developing an agent-first mobile application designed to integrate AI into daily communication and workflows with minimal prompting. Our mission is to eliminate misinformation and reduce coordination costs by bringing intelligence to everyday tasks.
As a Member of Technical Staff, you will own the lifecycle of ML models from development to production. You will work on practical problems involving search relevance, user behavior personalization, and system reliability.
What you'll do
- Build and improve ML models for classification, ranking, retrieval, and personalization
- Optimize classifier accuracy, precision, latency, and cost efficiency
- Enhance search relevance using embeddings, hybrid search, and reranking techniques
- Develop personalization models that learn from user behavior while preserving privacy
- Create evaluation pipelines, labeled datasets, and metrics for model quality
- Fine-tune existing models when sufficient data is available
- Maintain ML training, inference, and deployment pipelines
- Debug production issues and monitor model performance based on real-world feedback
What you'll need
- Strong foundation in machine learning and modern neural architectures
- Hands-on experience building and deploying ML models in production environments
- Experience with classification, ranking, retrieval, recommendation, or personalization systems
- Understanding of model evaluation, experimentation, and performance metrics
- Ability to write production-quality code and handle real-world datasets
- Skill in balancing model quality, latency, cost, and reliability
- Strong ownership mindset and problem-solving capabilities
Stack
- Python
- PyTorch
- JAX
- Embeddings
- Vector search
- Ranking models
Hiring process
- Technical team evaluates applications initially
- Schedule three to four interviews if there is a fit
- Interviews conducted virtually or onsite
- Prompt decision provided after final interview