Working with different types of data including computer logs, errors, natural languages, and other multimodal data
Work with state-of-the-art large language models
Document findings and approaches and present the results to the team
Optimize algorithms and prototypical solutions, such as hyperparameter search, model fine-tuning, etc.
Collaborate with the engineering team to implement (build, test, and deploy) the ML models
Create unit, integration, end-to-end, and/or performance tests.
Keep current with technology and industry developments and be on the lookout for new approaches and opportunities to integrate them into the existing solutions
Requirements & Skills:
Have strong analytical and problem-solving skills, and willing to dive deep to find creative solutions
Have deep experience around NLP and building custom LLM solutions
Have an entrepreneurial spirit and love to work in a fast-growing start-up environment, where you will meet like-minded people and celebrate successes
BS or MS/PhD degree in computer science, engineering, statistics, applied mathematics, or related quantitative discipline
Proficiency in programming languages commonly used in NLP such as Python
Strong knowledge of state-of-the-art pre-trained language models such as BART, Mistral, GPT, etc., along with expertise in fine-tuning these models for specific downstream tasks.
Experience in designing and implementing custom architectures and adaptations for large language models to optimize performance on specific tasks or domains.
Familiarity with transfer learning techniques and methodologies for fine-tuning pre-trained language models on diverse datasets.
Proficient in model evaluation and benchmarking methodologies for fine-tuned language models, including standard NLP evaluation metrics as well as task-specific metrics for tasks such as text generation, text classification, language modeling, etc.
Strong understanding of memory and computational constraints associated with fine-tuning large language models, along with experience in optimizing model inference and deployment for production environments.
Basic knowledge of agile software development (e.g. version control, kanban processes, and cloud deployment).
Team player. Comfort working in a dynamic group with open problems to solve.
