AI Researcher - Thorin

Title: AI Researcher

Location: San Francisco
Company: Thorin
Stage: Seed | Incubated at 8VC

About Thorin

Thorin is an applied AI company born out of 8VC’s Build program, focused on redefining how work gets done inside modern organizations. We build proactive, long-horizon AI agents that continuously observe work across Slack, email, meetings, and documents—and take action so nothing falls through the cracks.

Unlike traditional AI tools that wait for prompts or rely on structured systems, Thorin creates a real-time digital twin of how a business actually operates, enabling automation, coordination, and insight over months and years. Backed by 8VC and built with product vision from Joe Lonsdale, Thorin is building foundational infrastructure for agentic, AI-powered productivity.

With backing and product vision from Joe Lonsdale, Thorin is 8VC’s latest incubation, following the success of Affinity, Addepar, Saronic, and other unicorns.

About the Role

We’re hiring an AI Researcher who will help define and deliver the core research work that powers Thorin’s AI differentiator. You’ll work at the intersection of machine learning research, applied model development, and product impact, advancing models and methods that enable deeper understanding and automation of enterprise workflows.

This role blends rigorous research thinking with practical implementation — moving ideas from concept through experimentation and into usable components that shape Thorin’s product and capabilities.

What You’ll Do

Research & Innovation

• Develop novel machine learning research grounded in real-world product needs.

• Explore new model architectures, training paradigms, and evaluation techniques tailored to organizational workflow understanding and automation.

Model Development & Evaluation

• Design, implement, and test ML / AI methods that improve model performance on core tasks.

• Build and refine datasets, metrics, and evaluations that drive research progress and product impact.

Collaborate Across Teams

• Partner with engineering, product, and data teams to translate research insights into production-ready algorithms and features.

• Present research findings internally to inform strategic decisions and technical directions.

Prototyping & Deployment

• Build prototypes and proof-of-concept models that demonstrate feasibility and impact.

• Support iteration toward scalable, reliable systems suitable for integration with Thorin’s stack.

Stay at the Frontier

• Track and contribute to the latest developments in ML, deep learning, and AI systems.

• Represent Thorin’s research in the community when appropriate (writing, talks, shared benchmarks).

What You’ll Bring

• Advanced degree (Master’s/Ph.D.) in Machine Learning, Computer Science, Statistics, or a related field — or equivalent research experience.

• Strong foundation in deep learning, probabilistic models, and large-scale AI systems.

• Experience with modeling frameworks such as PyTorch or TensorFlow.

• Ability to bridge research and product engineering — designing experiments and implementing them with reliability in mind.

• Creative problem-solving skills and a track record of working in ambiguous, early-stage environments.

• Excellent communication — you can explain difficult concepts clearly and make data-driven recommendations.

Bonus Points For

• Publications or research contributions in relevant areas (NLP, representation learning, self-supervised learning, or workflow automation).

• Experience with LLMs, retrieval systems, or structured reasoning models.

• Prior work in enterprise AI, productivity automation, or human-machine collaboration.

• Familiarity with methods for fine-tuning, evaluation, or scaling of foundation models.

About the Team & Why Thorin

• Be among the first research hires and shape foundational AI direction.

• Work closely with a small, highly driven founding team, design partners, and directly with Joe Lonsdale, who remains deeply involved in product vision, strategy, and hiring

• Operate with high agency, rapid iteration, and direct customer exposure from day one

• Help define how a category-defining AI company goes to market from first principles

• Build with the backing of one of the strongest early-stage venture platforms in the industry