turbopuffer · May 2026
External engineering write-up from turbopuffer describing SID's RL training loop, search backend scaling, and why the model is trained as an iterative retrieval system.
Public company, workplace, funding, and market signals
Updated Jul 30, 2026
SID is a San Francisco-based AI research lab and software company focused on agentic retrieval and custom-data RAG. Its flagship product is SID-1, a retrieval model that searches iteratively and is sold as an API, via AWS Bedrock, and as a self-hosted model.
Primary product
SID-1 agentic retrieval model
Founded
2022
Headquarters
San Francisco, California, United States
Team size
1-10 employees
Work style
Onsite
Industry
Software Development
Sub-industry
AI research lab / retrieval software
Offices
0 jobs at SID
Check back later for new openings
Business model
Stage
pre-seed
Total raised
$500K
Latest round
Pre-Seed · May 2023
Latest amount
$500K
May 2023 · Y Combinator
Investors
Public hiring material portrays a small, research-heavy, high-bar team with startup intensity. The company emphasizes in-person work in San Francisco and Zürich, expects everyone to code, and says high-leverage research teams should stay small.
Work style
Onsite
Visa sponsorship
Limited
Compensation
The only public comp signal found is a YC job post for a Summer 2026 Research Intern at $5K-$10K per month; broader salary bands are not publicly disclosed.
Benefits
Pricing
Not publicly disclosed for commercial use; public materials describe API/self-hosted access and publish benchmark cost-per-question figures for SID-1.
Differentiators
Technology
Customers
Competitors
Estimated monthly visits
11.5K
turbopuffer · May 2026
External engineering write-up from turbopuffer describing SID's RL training loop, search backend scaling, and why the model is trained as an iterative retrieval system.
LinkedIn · Dec 2025
Founder announcement of SID-1's release, including performance claims versus embedding-only search and frontier-model agentic retrieval baselines.
SID AI · Dec 2025
Official launch of SID-1, an agentic retrieval model that is more accurate than embedding-only search and positioned as a subagent for larger LLMs.
SID AI · Dec 2025
Technical report detailing the model architecture, multi-turn RL training, benchmark results, and deployment characteristics.
Y Combinator · Jul 2023
YC launch page describing SID's original mission as a research lab for retrieval and early product positioning.