Traverse
Traverse is a San Francisco-based research data lab founded in 2025 that builds reinforcement learning environments and training data for frontier AI labs, targeting non-deterministic, taste-dependent work in domains like law, healthcare, sales, writing, and strategic decision-making.
- Company typePrivate
- Founded2025
- HeadquartersSan Francisco, United States
- Headcount1–10
- GTM typeB2B
- OfferingSoftware
What Traverse does
Traverse is a San Francisco-based research data lab founded in 2025 that builds reinforcement learning environments and training data for frontier AI labs. The company targets a specific unsolved problem in AI training: non-deterministic, taste-dependent work in domains such as law, healthcare, sales, writing, and strategic decision-making, where success depends on judgment, context, and reasoning rather than deterministic outcomes. Its core methodology involves observing real domain experts operating inside real environments and preserving the reasoning processes behind their decisions, then encoding that expertise into reward signals and evaluation criteria for model training.
The company's product offering is a single integrated platform — the RL Environments Platform — which is delivered through direct partnerships with frontier AI labs rather than via self-serve or marketplace channels. Traverse's go-to-market is enterprise field sales combined with an API-first integration approach, with revenue generated through subscription/recurring service contracts for access to proprietary RL environments and training data. Pricing is not publicly disclosed and is handled via direct outreach ([email protected]).
Traverse is backed by Y Combinator (lead investor in a $500K seed round) and has secured angel investments from OpenAI, Google DeepMind, Anthropic, and Meta, the same organizations that constitute its addressable customer base. The team is small (1-10 employees) and is actively hiring across research and engineering roles in San Francisco. The long-term stated ambition is to provide the training foundations for frontier models to perform and eventually surpass human white-collar work, accelerating progress toward artificial superintelligence.
Traverse firmographics
Firmographics- Name
- Traverse
- Legal name
- Traverse
- Website
- https://traverse.so
- Company type
- Private
- Founded year
- 2025
- Operating status
- Operating
- Headcount range
- 1–10 employees
- Short description
- Traverse is a San Francisco-based research data lab founded in 2025 that builds reinforcement learning environments and training data for frontier AI labs, targeting non-deterministic, taste-dependent work in domains like law, healthcare, sales, writing, and strategic decision-making.
- Ownership category
- akta.pro rank
Traverse industry classification
Industry- Product category
- AI Training Data and Reinforcement Learning Infrastructure
- NAICS
- Scientific Research and Development Services (5417), Other Scientific and Technical Consulting Services (54169)
- SIC
- Services-Computer Integrated Systems Design (7373), Services-Management Consulting Services (8742)
- akta.pro primary industry
- eLearning & Microlearning Module Development (EDAGAKAD)
- akta.pro secondary industry
- Learning Experience Platforms (LXP) & Learning Portals (EDAFAAAB)
Keywords
Where Traverse is headquartered
LocationHeadquarters
- HQ city
- San Francisco
- HQ country
- United States
- HQ region
- North America
Offices1 record
Markets served
Traverse business model
Business model- GTM type
- B2B
- Offering type
- Software
- Cost components
- Personnel, Technology or R&D, Operations, Marketing or Sales
Revenue model
- RL Environment Licensing to AI Labs: Providing specialized reinforcement learning environments and training data to frontier AI labs for integration into their post-training pipelines. Revenue is derived from service contracts with partner labs who pay for access to proprietary training environments and data.
Go-to-market motion2 records
Distribution channels1 record
Marketing channels4 records
Traverse product offering
Product offeringCore offering
Traverse is a research data lab that partners with frontier AI labs to produce reinforcement learning (RL) environments and training data for non-deterministic, subjective, and long-horizon tasks. Its work focuses on observing real experts operating inside real environments and preserving the reasoning process behind their decisions to create reward signals and evaluation criteria that help models develop taste and judgment across domains such as law, healthcare, sales, writing, and strategic decision-making. The offering is delivered as a single integrated research platform integrated directly into partner labs' post-training pipelines.
Product overview
Traverse operates as a single unified research platform focused on building reinforcement learning environments for frontier AI labs. The company specializes in creating training data for the non-deterministic, taste-dependent work that constitutes most of the economy — work where success depends on reasoning, context, and judgment rather than deterministic outcomes. The RL Environments Platform enables AI labs to train models on subjective and long-horizon tasks, with the goal of developing models capable of surpassing human white-collar work. There are no separate modules, add-ons, or sub-products — the offering is a single integrated research platform and methodology for encoding domain expertise into reward signals and evaluation criteria.
Differentiator
Problem solved
Functional benefit
Products and services
- RL Environments Platform An integrated research platform that produces reinforcement learning environments and training data for frontier AI labs, enabling models to develop taste, judgment, and reasoning on non-deterministic, subjective, and long-horizon tasks in domains such as law, healthcare, sales, writing, and strategic decision-making.
Companies that use Traverse
Customer profileNamed customers4 records
Segments1 record
Ideal customer profiles1 record
Traverse technology and API
TechnologyTechnology focussed Yes
API detail
- Has API
- No
- API docs
- API detail
Core technology
AI maturity
App detail
Feature4 records
Traverse partnerships and signals
Strategic signalPartnerships
One partnership is on record.
- Frontier AI LabscoreTraverse partners with frontier AI labs to produce training data required for models to develop taste and judgment. Labs integrate Traverse's RL environments into their post-training pipelines. Traverse works as a thought partner with these labs.
Scale indicators3 records
Recent moves5 records
Expansion highlights5 records
Traverse competitors and assessment
Company assessmentDirect peers
- Labelbox: Labelbox is a data labeling and annotation platform serving AI/ML teams with infrastructure for creating training datasets. It addresses the same data-centric AI training problem Traverse targets.
- Scale AI: Scale AI is the dominant provider of training data for AI labs, including RLHF data and evaluation infrastructure. It is the most direct competitor to Traverse, serving the same frontier AI lab customer base with adjacent data products.
- Invisible: Invisible (Invisible Technologies) operates AI training data operations, including RLHF labeling and expert workflows for AI labs. It is comparable to Traverse as a backend operations layer producing training data for frontier models.
- Turing: Turing provides AI engineering services and curated training data to frontier AI labs and enterprises. It overlaps with Traverse in supplying expert-driven training data and human-in-the-loop model improvement.
- Snorkel AI: Snorkel AI provides programmatic data labeling and weak supervision for AI training, enabling structured training data generation. It overlaps with Traverse's mission of producing scalable, high-quality training data for AI models.
- Datasaur: Datasaur provides data labeling and NLP annotation tools for AI/ML training pipelines. It addresses the same training data creation workflow that Traverse serves for AI labs.
- Surge AI: Surge AI specializes in high-quality RLHF and evaluation data for frontier AI models, including subjective/judgment-based tasks. It competes with Traverse in the same niche of expert-annotated training data for leading labs.
Broad incumbents
- Appen: Appen is a large, established provider of training data and annotation services for AI/ML, serving many of the same frontier lab customers. It is a broad incumbent with overlapping capability but less specialization in non-deterministic RL environments.
- CloudFactory: CloudFactory provides managed AI data annotation and labeling workforces for ML teams. It is comparable to Traverse as a scaled data operations provider, though with broader and more general-purpose offerings.
Emerging players
- Generally Intelligent: Generally Intelligent is an AI research lab focused on human-like general intelligence, with comparable frontier research ambitions. It overlaps with Traverse in mission and culture (small research-first team targeting AGI) but focuses on model research rather than data infrastructure.
Market position
Strengths5 records
Weaknesses5 records
Competitive moat4 records
Key highlights6 records
Customer concentration
Traverse social profiles
Digital presenceTraverse financial estimates
Financial estimateRevenue estimate
Valuation estimate
Traverse leadership team
Management profileNumber of profiles
Profiles2 records
Traverse funding detail
Funding detailFunding overview
Funding rounds1 record
Investors3 records
Funding detail is available on the Subscription and Enterprise plan.Contact sales →
Traverse M&A and investment
M&A and investmentM&A
Investments
M&A and investment is available on the Subscription and Enterprise plan.Contact sales →
Frequently asked questions about Traverse
What does Traverse do?
Traverse is a research data lab that partners with frontier AI labs to produce reinforcement learning (RL) environments and training data for non-deterministic, subjective, and long-horizon tasks. Its work focuses on observing real experts operating inside real environments and preserving the reasoning process behind their decisions to create reward signals and evaluation criteria that help models develop taste and judgment across domains such as law, healthcare, sales, writing, and strategic decision-making. The offering is delivered as a single integrated research platform integrated directly into partner labs' post-training pipelines.
Is Traverse a public or private company?
Traverse is a private company. It is classified as venture growth investor backed and is currently operating.
When was Traverse founded?
Traverse was founded in 2025. It employs 1 to 10 people.
Where is Traverse based?
Traverse is headquartered in San Francisco, United States, in the North America region.
How does Traverse make money?
One revenue line is on record: RL Environment Licensing to AI Labs.
Who are Traverse's main competitors?
Direct peers on record are Labelbox, Scale AI, Invisible, Turing, Snorkel AI, Datasaur and Surge AI. Broad incumbents are Appen and CloudFactory. Generally Intelligent is listed as an emerging player.
Does Traverse have an API?
No public API is recorded for Traverse.
What industry is Traverse in?
Traverse's product category is AI Training Data and Reinforcement Learning Infrastructure. Its primary akta.pro industry code is EDAGAKAD, eLearning & Microlearning Module Development, with a secondary code of EDAFAAAB, Learning Experience Platforms (LXP) & Learning Portals. Its NAICS code is 5417 and its SIC code is 7373.