Senior Data Scientist II

Major Responsibilities:

• Work closely with team of AI engineers to design, build, and serve LLMs to solve complex business challenges using Azure (CPU & GPU environments)
• Researching & implementing state of the art LLM techniques including pre-training, fine-tuning, preference alignment, and deployment while also focusing on prompt engineering and generative AI more broadly
• Heavily focusing on developing novel data sets that enable LLMs to perform new tasks as well as tooling/platforming to collect these samples at scale. You will need strong python data fundamentals coupled with a software mindset for making data processing and collection pipelines repeatable, scalable, and high quality.
• Ensure high quality code that meets business objectives, quality standards and development guidelines.
• Building reusable pipelines, processes, and tools to streamline LLM and generative AI workflows
• Manage project stakeholder expectations and issue communications on progress.
• React to shifting priorities without compromising deadlines and momentum.

Qualifications:

• Must Have:

• Nice To Have:

• 3 + years’ experience in AI/ML. 
• Experience in LLM Engineering – pretraining, post-training /alignment
• Deep expertise in writing and reviewing production code in Python.
• Understanding the development lifecycle for LLMs— developing data sets for pre-training, instruction tuning, and preference alignment alongside the modelling techniques for each stage and LLM deployment is as MAJOR plus.
• Multi-disciplinary approach to problem solving, including excellent interpersonal and communication skills (written and verbal). This includes crisply talking about technical solutions while being able to collaborate with business architects effectively.
• Strong knowledge of LLM frameworks and libraries (such as transformers, trl, deepspeed, PyTorch), and exposure to various ML techniques and their practical implementation in production at large scale.
• Experience on distributed, high throughput and low latency architectures
• Strong fundamentals in NLP techniques for text representation, semantic extraction techniques, data structures and modeling.
• Experience building software on top of major container technology (Kubernetes, Docker etc.)
• Experience defining system architectures and exploring technical feasibility tradeoffs is a huge plus
• Familiarity with end-to-end application development using full stack is a plus.
• Experience in P&C insurance is a plus.