Senior Data Scientist - LLM

Overview: 

Spotter, named one of TIME100's Most Influential Companies this year, empowers top YouTube creators to accelerate their business and unleash their full creative potential by giving them access to the capital, knowledge, and community they need to succeed at scale. As the top provider of creator-friendly growth capital, Spotter tailors our investments to meet the unique needs of each creator we partner with, giving them the freedom to create without compromise.

Creators are free to reinvest their funds however they choose, from hiring a team, to building their own production studios, and everything in between, all while maintaining total control over their catalogs, their channels, and their future earnings. In addition to funding, Spotter provides creators with in-depth data insights into the performance of their existing content, enabling them to leverage the full value of their library, as well as the value of future uploads and how they can improve performance in the future.

Featured in Forbes, Fast Company, Variety, Axios, and more, Spotter has already deployed over $850 million to YouTube creators to reinvest in themselves and accelerate their growth. Spotter has licensed content that consists of over 725,000 videos, which generate 88 billion monthly watch-time minutes. With our curated premium video catalog, we deliver a unique scaled media solution to Advertisers and Ad Agencies that is transparent, efficient, and 100% brand safe.

We’re looking for a talented and intensely curious LLM specialist and deep learning modeler that is nimble and focused with a startup mentality: You want your models and code to ship and improve the lives of real customers, not rot in a research department. You’ll be fine-tuning and serving various models, but you’ll also obsess over learnings in the industry and apply them in your daily work. Like for all of us, the outcome matters more to you than how we get there. You get a kick out of serving tools that improve and accelerate the workflows of the top YouTube creators in the world! If that all sounds like you, keep on reading.

What You’ll Do

• You’ll fine-tune and serve LLMs and LLM-based pipelines to help build best-in-class workflow tools for the world’s top YouTube creators, and you’ll ship updates and new features on a weekly basis
• You’ll be part of a small but mighty and highly motivated team (~15 people strong) that operates like a startup within a startup–with all the chaos that comes with that–but also real camaraderie, diverse backgrounds, skills and opinions and plenty of fun.
• You’ll learn and grow every day - from the industry, your peers and most importantly, your customer zero - the creators!

Responsibilities

• Fine-tune pre-trained language models for generative AI applications.
• Apply, discover and research techniques to optimize existing LLM training and serving, as well as improving the model quality.
• Develop and deploy the LLMs for a variety of content generation tasks.
• Share your discoveries, learnings and results with other modelers in the company working on related projects.
• Continuously research and stay up-to-date with recent advancements in NLP and large language models, applying novel techniques and methodologies to improve our models.
• Read all the newsletters and obsess over new industry learnings, startups utilizing LLMs in novel ways, and other ways the real world is utilizing LLMs.
• Conduct experiments and benchmarking to assess the performance of various model architectures and optimize hyperparameters.
• Troubleshoot and resolve any issues arising during model training and deployment.
• Work with product teams to ensure smooth product delivery with high standards, prioritize short-term returns and long-term growth.

Who You Are

• Master’s degree in Machine Learning, Computer Science, Electrical Engineering, Statistics, Applied Math, or other quantitative fields. 
• Hands-on experience in developing models, in optimizing model training and tuning and in deploying Large Language Models such as GPT, LLaMA, BERT, or Transformer-based architectures with strong knowledge in tokenization and embeddings.
• 2+ plus years of experience in developing deep learning models as a Data Scientist and experience in building products backed up ML /AI
• Experience with transfer learning, meta learning and predictive modeling.
• Expertise in visualizing and manipulating big datasets.
• Strong analytical and problem-solving skills and 
• Intensely curious to learn how AGI is evolving day in and day out.
• Excited about creators 
• Excellent oral and written communication skills.
<