At FurtherAI, we’re creating AI Agents to automate tasks traditionally requiring human expertise. As the Founding Applied Research Engineer, you’ll work on developing and productizing product-focused solutions, leveraging Large Language Models (LLMs) and Vision Language Models (VLMs) to solve real-world challenges and push the boundaries of what AI can achieve.

LLMs can solve 80% of a problem, but the leap from 80% → 100%is where true innovation happens. Your mission: capture the peaks of intelligence while ensuring the reliability and consistency of human-level performance.

This is an  in-person role based in  San Francisco, CA .

• Build an agent swarm to automatically triage inbox
• Accurately extract structured and unstructured data from any type of document.
• Design validator systems to detect AI failures to escalate them to human reviewers.
  

• Passion for iterating fast and building things - if you’re exceptional in this area, other requirements are optional.
• Experiment with novel techniques to improve the accuracy and reliability of LLMs and VLMs.
• Expertise in fine-tuning existing foundation models.
• Proficiency in creating and curating training and evaluation datasets
• Deploying LLMs in production environments
• Staying up-to-date with the latest advancements and trends in AI.
• Bonus: Publishing in leading AI/ML conferences (e.g., NeurIPS, CVPR, ACL).
  

At FurtherAI, we take immense pride in our work, operate with a strong sense of ownership, and are relentless in our pursuit of excellence. Every member of our team works extremely hard, driven by the common goal of building something transformative. If you’re not ready to bring this level of commitment and ownership to the table, this role may not be the right fit for you.

• Build new product experiences that use AI agents as centerpiece
• Work with cutting-edge LLM and VLM models
• Be one of the first employees in a fast-paced startup
• Start your own company in the future (66% of our team is ex-founders!)
  

At FurtherAI, we are building a workforce of AI Teammates to automate insurance workflows. These AI Teammates can practically automate any insurance workflow involving processing unstructured documents, data entry into internal systems or web portals, or even making phone calls.

Our core mission is to answer a critical question: Can AI be made as reliable, adaptive, and continuously learning as a new human employee?

We have successfully raised a seed round from top-tier investors, including Y Combinator, South Park Commons, and Converge VC (full announcement soon). Our founding team features a second time founder and a language modeling scientist from Apple, who have known each other for over 12 years.