ClickHouse · Jul 2026
Partner/customer story explaining how Trainy moved Pluto's metadata layer to ClickHouse Managed Postgres, improving page loads up to 3x and reducing operating costs.
Infrastructure for managing GPU clusters for training/serving.
Public company, workplace, funding, and market signals
Updated Jul 29, 2026
Trainy is a YC-backed San Francisco AI infrastructure startup that helps teams schedule, monitor, and optimize GPU training and inference workloads across customer-owned cloud or bare-metal infrastructure. Its two core products are Konduktor (GPU cluster/job scheduling and management) and Pluto (open-source, Neptune/W&B-compatible experiment tracking).
Primary product
Konduktor and Pluto
Founded
2023
Headquarters
San Francisco, California, United States
Team size
1-10
Industry
Software Development
Sub-industry
AI infrastructure / GPU cluster management
Offices
0 jobs at Trainy
Check back later for new openings
Business model
Stage
Pre-Seed
Total raised
$500K
Latest round
Pre-Seed · Sep 2023
Latest amount
$500K
Sep 2023 · Y Combinator
Investors
Technical, infrastructure-first, and open-source oriented; public writing emphasizes reliability, observability, performance, and cost efficiency for ML teams.
Pricing
Konduktor uses custom quote / platform fee plus per-node pricing; Pluto offers a free hosted tier and a $250/seat/month hosted Pro tier, with free self-hosting.
Differentiators
Technology
Customers
Competitors
Estimated revenue
about $300K-$400K/year (public estimate)
Estimated monthly visits
3.5K
ClickHouse · Jul 2026
Partner/customer story explaining how Trainy moved Pluto's metadata layer to ClickHouse Managed Postgres, improving page loads up to 3x and reducing operating costs.
Trainy blog · Feb 2026
Trainy argues Neptune's shutdown exposed the value of fast experiment tracking and positions Pluto as a migration path for Neptune users.
Fondo · Feb 2026
External write-up on Trainy's Pluto launch, including the Neptune migration angle and hosted/self-hosted pricing details.
Trainy blog · Mar 2025
Technical post focused on validating GPU network fabric before multi-node training, reinforcing Trainy's infrastructure-first positioning.
Trainy blog · Feb 2025
Technical post arguing that GPU utilization alone is a poor indicator and advocating more meaningful performance metrics.