Deep Learning Engineer (NLP / CV) in Anadea

14 серпня 2024

Deep Learning Engineer (NLP / CV)

віддалено

Join our dynamic team as a Data Scientist/DL Engineer and drive innovation with cutting-edge ML/DL/AI solutions across diverse business domains.

Who We Are Looking For:

Must have:

• 1+ year of experience as ML/DL Engineer
• Profound understanding of ML/DL fundamentals
• Proficiency in Pytorch or Tensorflow — expertise in training models, identifying and resolving issues and optimizing model performance.
• Understanding of deep neural networks — what loss function to use, how to fix exploding gradients, why do we need regularization
• Proven experience with either CV or NLP tasks and architectures — understanding the difference between T5 and BERT or ResNet/EfficientNet is required.

Nice to have:

• Experience in containerizing DL pipelines (Docker).
• Commitment to high-quality coding: docstrings, type annotations, understanding of tests.
• Commitment to tracking your work: experience with DVC, MLFlow, Tensorboard or similar for tracking datasets and experiments.
• Advanced English proficiency, strong communication skills, ability to easily and clearly articulate technical concepts and effectively engage with the clients.
• Experience with Generative AI: LLMs, Image Generation models, understanding adapters, RAG

What we offer:

• Variety of Projects: Engage with different areas like CV, NLP, and Tabular data, ensuring no two days are the same.
• Top-Tier hardware: servers with A6000/RTX 4090, access to high-end GPU instances on cloud.
• Environment: work with an experienced team passionate about the industry.
• Direct involvement with clients, opportunity to influence projects with your ideas and solutions.
• Opportunities for business trips, participation in conferences, PR of our work and your contributions.

Interview Process:

• HR interview
• Introduction (30-45 minutes).
• Take-home assignment (6-8 hours).
• Solution discussion (60 minutes).
• Offer

How to Apply:

CV is essential, links to GitHub/Kaggle/your outstanding papers and projects are very welcome and highly encouraged