Fulcrum is an agentic debugger for AI systems, building red-teaming agents that run experiments to diagnose why AI agents fail. The platform plugs into environments, agent traces, and agent source code to find root causes, uncover environment bugs, detect reward hacking, and turn investigations into explorable diagnostic reports. The founders met at MIT doing LLM research and developed methods for scalably generating software environments of arbitrary difficulty. Fulcrum was co-founded by Kaivalya Hariharan and Uzay Girit, both MIT graduates, and went through Y Combinator's Summer 2025 batch.
Fulcrum
ActiveThe agentic debugger for AI systems
Total raised
$500K
1 round
Stage
Seed
Jan 2025
Team
1-10
since 2025
Pricing
—
Founded
2025
San Francisco, United States
Agent-ready
—
Agentic debugger that diagnoses why AI agents fail
Red-teaming agents that run experiments to find root causes
Plugs into agent environments, traces, and source code
Detection of environment bugs affecting agent behavior
Reward-hacking detection in AI systems
Root-cause analysis of agent failures
Explorable diagnostic reports from investigations
Methods for scalably generating software environments of varied difficulty
12/100
Early
MCP server
Public API
Webhooks
OAuth 2.0
SDKs
No public agent surfaces detected yet.
Jan 2025 Seed $500K ● Y Combinator
Capital network
$500K raised ·1 backer·10 network links
- Backers1
- Shared portfoliocompanies these backers also fund
- Extended networkfunds that co-invest alongside them
Findem
Expert-labeled talent dataset for AI-powered hiring
AI AnalyticsAI Recruiting
Pearl Health
AI value-based care platform for primary care providers
AI AnalyticsAI Healthcare
Sigma Computing
AI Analytics
Endgame
AI answer engine for enterprise sales teams
AI ProductivityAI Sales
Peregrine Technologies
Decision-support and data integration platform for public safety agencies
AI AnalyticsAI Governance
Sylvera
Carbon credit ratings and market intelligence powered by AI and geospatial data
AI AnalyticsAI Climate
- What does Fulcrum do?
- Fulcrum is an agentic debugger that uses red-teaming agents to run experiments and diagnose why AI agents fail, plugging into environments, traces, and source code.
- What kinds of issues can it find?
- It can find root causes of failures, uncover environment bugs, and detect reward hacking, turning investigations into explorable diagnostic reports.
- Who is Fulcrum for?
- It targets teams building and operating AI agents who need to understand and fix complex, hard-to-diagnose failures.
- Who founded Fulcrum?
- It was co-founded by Kaivalya Hariharan and Uzay Girit, both MIT graduates who met doing LLM research, and went through Y Combinator's Summer 2025 batch.
Discussion
Sign in to join the discussion.
Sign inExplore more around Fulcrum
Contextual paths to related AI startups, deals and rankings.