rekursiv.ai GTM brain

rekursiv.ai offers an autonomous research cockpit that runs AI scientist teams for technical researchers.

category

Autonomous AI research agents

market profile

  • text: Autonomous research systems seek to generate hypotheses, run experiments and analyze results, but independent verification of findings remains important.
  • text: rekursiv.ai presents evidence from machine learning, benchmark optimization and factory scheduling, and offers access by email.

personas

  • title: Head of machine learning research
    pains:
    • Limited researcher time constrains the number of viable hypotheses tested.
    • Model quality improvements must justify experimentation and compute costs.
  • title: Research scientist or research engineer
    pains:
    • Experiment setup, running evaluations and interpreting results compete with idea generation.
    • Surprising results need reproducible evidence before the team can trust them.
  • title: Industrial optimization or applied research lead
    pains:
    • Scheduling and combinatorial optimization require domain-specific experimentation.
    • Expert-designed solvers can be expensive to improve and benchmark across problem instances.

icp

  • text: Research-intensive AI labs and technology companies with quantitative ML benchmarks and the ability to evaluate novel methods; account size is not specified.
  • text: Applied R&D teams with well-defined computational optimization problems and representative test instances, such as scheduling.
  • text: Early-access buyers willing to submit a research problem and validate results with their own evaluation process.

message pillars

  • text: Turn research problems into autonomous experimentation.
  • text: Make research claims traceable.
  • text: Scale idea generation and experimental work.
  • text: Show research-native credibility.

objections

  • objection: Are agent discoveries actually novel and reproducible?
    response: Request experiment traces, baselines and held-out evaluations, then independently reproduce the result before adoption.
  • objection: Will it help our research rather than only benchmark tasks?
    response: Run a narrowly scoped pilot against a known baseline with agreed compute limits and success metrics.
  • objection: What happens to proprietary research data and IP?
    response: Request written details on data handling, model access and ownership before sharing sensitive data.

competitors

  • name: Sakana AI
    website: https://sakana.ai/ai-scientist-jp/
    why: Its AI Scientist system automates parts of the research lifecycle.
  • name: Edison Scientific
    why: Commercializes FutureHouse scientific research agents including Kosmos.
  • name: Weco AI
    why: Builds autonomous research tooling listed in the automated research landscape.

influencers

  • name: Joshua V. Dillon
    url: https://rekursiv.ai/josh/
    why: Co-founder and former Google and DeepMind researcher.
  • name: Dan Kondratyuk
    url: https://dankondratyuk.com/
    why: Co-founder and machine learning researcher writing about autonomous science.

keywords

  • autonomous AI scientist
  • AI research agents
  • automated machine learning research
  • AI hypothesis generation experiments
  • AI experiment optimization
  • autonomous algorithm discovery
  • AI combinatorial optimization

voice

  • research-led and technical
  • ambitious and discovery-oriented
  • evidence-focused

signals

  • text: A lab is investing in ML benchmark improvement and runs many repeatable experiments.
  • text: An industrial R&D team has a measurable optimization task and a baseline solver.
  • text: A research team is actively evaluating agent-based research automation or compute efficiency.

pricing

model: unknown

motion

sales_led

channels

  • research blog
  • GitHub
  • YC directory
  • LinkedIn
  • research community

confidence

medium

Founders

  • name: Joshua Dillon
    title: Founder
    role: Founder
    bio: Ex-Google, Ex-Deepmind, Ex-LumaAI staff research scientist for a combined total of 14 years. Led foundational model pre-training at Luma AI. Creator of TensorFlow Probability (1M+ downloads / month), built Veo's first prototype, contributor to Gemini and VideoPoet (ICML 2024 Best Paper). 11,600+ citations. Now co-founder of rekursiv.ai, automating the science of ML research.
  • name: Dan Kondratyuk
    title: Founder
    role: Founder
    bio: Ex-Google, Ex-LumaAI scientist for a combined total of 7 years. First author of VideoPoet (ICML 2024 Best Paper), led the World Models team at Luma, and built large-scale multimodal LLMs and diffusion models for image, video, audio, and text. Now co-founder of rekursiv.ai, automating the science of ML research.

Back to YC companies