Job Description
We are a US-based outsource software development company that has been delivering exceptional software experience to our clients since 2011, helping technology companies to become industry leaders.
Over the past few years, we’ve been hiring specialists all over the world while our main development centers were in Ukraine. Now, we keep expanding and start growing our centers in different parts of the world. Dev.Pro is open to hire specialists from other countries as well as Ukrainians who live outside of Ukraine now. We stand with Ukraine and keep supporting our people by offering a friendly remote environment while adhering to the values of democracy, human rights, and state sovereignty.
We are looking for a hands-on Data Scientist or Applied Machine Learning Engineer to work closely with a team of researchers building and refining AI agents and training environments. This role sits at the intersection of research and execution. The ideal candidate is someone who is comfortable taking partially formed research ideas or prototypes and turning them into robust, working implementations that can be evaluated, debugged, and ultimately productionized.
This is not a traditional software engineering role and not a purely academic research role. We are explicitly looking for a strong data science profile with deep comfort working directly with models, training workflows, evaluation logic, and subtle implementation details.
Key Responsibilities
• Work directly from researcher-written task descriptions that define what a model is being trained to do,
including the design and implementation of graders and evaluation logic
• Implement, refine, and debug training and evaluation pipelines using PyTorch (and/or JAX where
applicable)
• Investigate and diagnose unexpected model behaviors by inspecting training dynamics, gradients,
loss functions, and implementation details
• Translate unpolished research prototypes into clean, correct, and maintainable implementations
• Identify and fix conceptual or mathematical errors in model implementations, including incorrect
indexing, off-by-one errors, incorrect tensor operations, or flawed assumptions
• Adjust and improve prompts, graders, reward functions, or training logic to correct failure modes
• Collaborate asynchronously with researchers with minimal oversight
• Perform extensive QA on tasks and training setups to ensure correctness and reliability
- Strong background in data science, machine learning, or applied ML
- Hands-on experience with PyTorch; JAX experience is a strong plus
- Comfort working with existing model implementations such as Hugging Face Transformers
- Strong understanding of gradient-based optimization, backpropagation, loss functions, and neural network architectures
- Ability to debug both code-level and conceptual modeling issues
- Experience working in fast-paced startup or research-adjacent environments
- Strong written and verbal communication skills in English
🌴🏄🏻 Maintain a healthy work-life balance with 30 paid rest days per year to use as holidays/vacation/other on the desired and requested dates
💎📚 Grow by learning: you have a chance to shift your career path, access English lessons, DP University educational flows, and other educational initiatives
💵💰 Make sure the compensation matches your efforts: we provide payments in US dollars and cover overtime hours
💃🏼🕺 Join fun online activities and team-building events
👩💻🖥 Get continuous remote HR and payroll support
