Machine Learning Engineer (Generative AI) (вакансія неактивна)

Galactica Labs is looking for a machine learning engineer to work with our long term legal technology client that is developing document drafting and workflow automation products.

We are looking for a brilliant engineer to explore cutting-edge large language model (LLM) applications in an applied data science role. You will be working to tune and train models for a variety of legal technology use cases. You will also work to extend and upgrade existing NLP/NLG features in production.

We have strong belief that LLMs will revolutionize industries and that we are uniquely positioned to take advantage of this technology in our niche of legal technology. You will be the first dedicated subject matter expert on LLMs and we will lean on you to pioneer LLM applications.

What this role is:

• Moving quickly to prototype and iterate on new LLM applications.
• Going end to end from data to model deployment
• Being able to evaluate different approaches and explain tradeoffs
• Staying on top of emerging trends in AI, LLMs and NLP
• Being a product thinker in the sense of reformulating existing problem statements to a form that AI/ML can tackle.
• Working independently, taking initiative, setting up experiments and demonstrating progress through principled use of metrics.

Requirements:

• Strong programming skills, with experience in Python and ML libraries (Pytorch, Tensorflow)
• Proven track record of ideation, designing and building prototypes as well as demonstrating, documenting, and sharing the learnings
• 3+ years of experience with building and deploying ML applications
• Experience with natural language processing (NLP)and large language models such as BERT, GPT, or Transformers
• Experience with NLP packages such as SpaCy, CoreNLP, OpenNLP, NLTK
• Strong problem-solving and communication skills
• Bachelor’s or Master’s degree in Computer Science or a related field
• Excellent interpersonal and English verbal and written communication skills
• Flexible working hours (must have some overlap with US ET time zone for team calls)