# Experience

Canonical: https://www.spencerhong.com/work/experience

## Experience

### General Intelligence Company of New York

Founding Researcher · Nov 2025 – Present

Leading evaluation, telemetry, and self-improvement research for long-running agents that operate businesses autonomously. Built the agent framework behind Cofounder 1 and Cofounder 2.

- Grew the company to 13 people, raised $10M+, and created products that reached 15,000+ monthly active users.
- Led evaluation, telemetry, and self-improvement research for long-running agents designed to operate businesses autonomously.
- Built the in-house agent framework spanning model integrations, agent harnesses, and infrastructure for Cofounder 1 and Cofounder 2.
- Trained models to improve unverifiable “taste” in brand design and go-to-market, using reinforcement learning with expert-authored rubrics and continued pretraining on new data streams.
- Led collaborations with frontier AI labs and academic researchers, including CEO-Bench, on benchmarking and data sourcing.

[General Intelligence Company of New York](https://www.generalintelligencecompany.com)

### Capital One

Principal Data Scientist · May 2024 – Nov 2025

Built and evaluated retrieval-augmented generation and agentic systems for bank support and financial investigations, including an anti-money-laundering system with 97% agreement with human investigators.

- Developed and evaluated retrieval-augmented generation systems to reduce handling time for hundreds of bank support agents.
- Built an uncertainty-based system to detect hallucinations in AI agent outputs at inference time.
- Managed researchers establishing a fault attribution framework to diagnose failures in multicomponent AI agent systems.
- Built an anti-money-laundering agentic system that achieved 97% agreement with human investigators.

### Northwestern University

Ph.D. Candidate and Ryan Fellow · 2020 – 2025

Led a multidisciplinary effort to study NIH funding through more than two million pages of archival documents. Trained multimodal models for document analysis and built an AI scientist framework integrating 120+ data sources.

- Ryan Fellow from the International Institute for Nanotechnology.
- First Ph.D. student in NIH history to present to the National Advisory Council for Human Genome Research.
- Designed, implemented, and led efforts to uncover fundamental funding mechanisms at the National Institutes of Health through an archive containing more than two million pages of unstructured, privileged documents.
- Trained multimodal deep learning models for PII redaction, entity disambiguation, handwriting separation, scan segmentation, document categorization, and topic mapping.
- Built an AI scientist framework in 2024, integrating 120+ data sources to identify understudied genes using natural language.
- Led a team of ethicists, archivists, historians, and engineers for the first data-driven investigation of a major scientific funding agency in history.

## Navigate

- [Overview](https://www.spencerhong.com/)
- [Research](https://www.spencerhong.com/work/research)
- [Music](https://www.spencerhong.com/music)
- [About](https://www.spencerhong.com/about)
- [Contact](https://www.spencerhong.com/contact)
- [Privacy](https://www.spencerhong.com/privacy)
- [Blog](https://www.spencerhong.com/blog)
- [On Agentic Memory](https://www.spencerhong.com/blog/on-agentic-memory)
- [AI agents](https://www.spencerhong.com/agents)