REBUY - Senior Applied Research Scientist

About Us:

Rebuy is SaaS application that is redefining data-powered shopping experiences with a new breed of personalization and merchandising solutions, which turn intelligence into revenue generating actions.

We offer onsite solutions, marketing tools and automation, RESTful API, JavaScript Library, and more to help super charge high growth brands.

Rebuy is a fully remote company and trusted by some of the fastest growing brands in DTC such as Aviator Nation, Liquid Death, Magic Spoon, Blenders, Laird Superfoods, Primal Kitchen and many more.

We value ownership, self-motivation, and empathy. Each team member plays a significant role in the company, and is helping to shape the future of intelligent commerce. We build strong and lasting partnerships with all of our merchants and partners, provide a safe and fun work environment, and truly love what we do.

About the role:

We are looking for a Senior Applied Research Scientist to join the AI/ML team at Rebuy. This role will be focused upon applied research that explores the use of large language models for e-commerce. At Rebuy, the AI/ML team leverages cutting edge deep/machine learning techniques to build personalized search and recommendation tools. Recently, we have invested heavily into the creation of e-commerce-focused dialogue tools (powered by large language models) that provide an alternative medium for product discovery beyond search and recommendations. As a Senior Research Scientist on the AI/ML team, you will do applied research that combines large language models (both APIs and open-source models) with a massive e-commerce dataset to create useful tools and dialogue systems for both e-commerce merchants and the shoppers who visit their store. The primary goal of this research is to create practical tools that can be used in and integrated with the Rebuy platform. However, the AI/ML team is interested in interfacing with the broader research community by publishing open-source tools, models, and papers whenever possible. The applied research conducted as part of this position will directly influence the company’s technical direction on AI-related initiatives as we continue to expand our e-commerce product offerings related to chat, search, and recommendations.

What you’ll do:

• Explore scalable and personalized applications of language models (both APIs and open-source models) on e-commerce stores via advanced prompting techniques, complex agent/router systems, retrieval/injection of relevant information, tool integrations, and more.
• Explore different avenues of training and alignment (e.g., pre-training, instruction tuning, SFT/RLHF, etc.) for open-source language models. Identify effective techniques for using massive amounts of commerce-centric data to create better, specialized language models for e-commerce applications.
• Implement systems for effectively benchmarking, measuring, and monitoring language model performance across a variety of different applications. Explore ways of improving the reliability of systems that leverage language models “in the wild”.
• Write technical blogs and publications that overview notable advancements. Publish and promote open-source models/tools on platforms like HuggingFace. Generally interact with the broader research community to help advance the frontiers of open-source research on language models.
• Interact frequently with other members of the AI/ML team to discuss research, ideas, and ML-related solutions to company problems. Identify opportunities to implement and deploy useful applications of AI.

Requirements:

• Willingness (and excitement) to be an active and impactful member of a close team of AI/ML experts that work every day to make cool ideas a reality.
• The ideal candidate:
  • Stays up-to-date with trends/research in AI/ML.
  • Has excellent listening and communication skills.
  • Effectively manages their time between competing priorities.
  • Is willing to work both independently and in collaboration with others to meet team objectives.
• Work/research experience:
  • Graduate Degree or equivalent of work experience.
  • 5+ years of experience with research in AI/ML.
• Technical proficiencies:
  • Familiar with the landscape of open-source language models (e.g., LLaMA-2, MPT, Falcon, etc.), paid language model APIs (e.g., Claude-2, PaLM-2, GPT-3.5/4, etc.), and language model research in general.
  • Experience with pre-training or fine-tuning large language models via open-source code repositories (e.g., LLM foundry, TRL, PEFT, etc.), including end-to-end or parameter-efficient training techniques (LoRA/QLoRA).
  • Understanding of prompt engineering, including both basic (e..g, zero/few-shot and instruction prompting) and more advanced (e.g., chain/tree of thought prompting, prompt ensembles, self-consistency, etc.) techniques.
  • Understanding of and practical experience with relevant distributed training frameworks such as FSDP. General familiarity with distributed training of large neural networks on cloud compute infrastructure.

Nice to have:

• Ability to clearly communicate what you are working on, how it works, and how it solves a tangible problem faced by the company.
• Experience with MLOps, scalable computing, and data infrastructure tools (e.g., Kubernetes, Prefect, Dask, MLFlow, etc.).
• Experience with deploying neural networks into production environments.
• Understanding of search and recommendation systems.
• Remote work experience.
• Startup experience.