Algorithmic SuperIntelligence Labs
Algorithmic SuperIntelligence Labs is an open-source research initiative associated with AFGI that aims to invent the next generation of algorithms. It combines large AI models, a discovery engine, and domain expertise to deploy AI-discovered algorithms that are faster, safer, and more reliable.
At a Glance
- University and academic researchers
- Technology companies and AI infrastructure teams
- Enterprise finance organizations
- GPU and hardware optimization teams
- +1 more
AI Tools by Algorithmic SuperIntelligence Labs
(2)
OpenEvolve
Open Source Evolutionary Coding Agent
OptiLLM
OpenAI Compatible Inference Proxy
Discussions
No discussions yet
Be the first to start a discussion about Algorithmic SuperIntelligence Labs
Latest News
How IBM Used OpenEvolve to Discover Quantum Error Correction Codes
AFGI Research to develop AI Lab for enterprise finance; article describes AFGI as running Algorithmic Superintelligence Labs
How AMD Used OpenEvolve to Optimise GPU Kernels
OpenEvolve for Quasi-Monte Carlo Design
Products & Services
An open-source implementation of AlphaEvolve and an evolutionary coding agent that uses LLMs, population-based evolution, island-based search, evaluation pipelines, and Pareto optimization to discover and optimize algorithms and programs.
An OpenAI API-compatible optimizing inference proxy that applies 20+ inference-time techniques to improve LLM accuracy and performance without model training or fine-tuning.
A Hugging Face Space that uses OpenEvolve to optimize prompts.
Market Position
The lab positions its software as an open-source alternative and complement to proprietary algorithm-discovery systems such as Google's AlphaEvolve, using LLM-driven evolutionary search rather than only manual optimization or fixed heuristics. OptiLLM competes in inference-time optimization by improving existing models through additional test-time computation rather than retraining; the lab's applied-research positioning is also comparable to small research organizations such as Sakana AI, with AFGI emphasizing enterprise finance.
Leadership
Founders
Ritik Vijayvergiya
AFGI Research founder; IIT Delhi alumnus and former Vice President at WorldQuant. The Hindu BusinessLine identifies AFGI as the firm that runs Algorithmic Superintelligence Labs.
Executive Team
Ritik Vijayvergiya
Founder of AFGI Research, the organization reported to run Algorithmic Superintelligence Labs
IIT Delhi alumnus and former Vice President at WorldQuant.
Asankhaya Sharma
Creator of OpenEvolve and OptiLLM
GitHub profile identifies him as the creator of OptiLLM and OpenEvolve; his profile lists Singapore and CTO of Patched Codes.
Founding Story
The initiative was created around the idea that large language models can be treated as a unit of compute and combined with automated discovery and domain expertise to discover algorithms beyond conventional hand-designed approaches. AFGI describes the lab as part of its open-source research initiative and broader lab-as-a-service effort for enterprise finance.
Business Model
Revenue Model
AFGI Research describes a lab-as-a-service model in which teams work directly with client organizations to design, build, and continuously deploy custom AI systems for enterprise finance. The Algorithmic SuperIntelligence projects themselves are presented as open-source research software.
Target Markets
- University and academic researchers
- Technology companies and AI infrastructure teams
- Enterprise finance organizations
- GPU and hardware optimization teams
- Scientific and mathematical research groups
- Algorithm and program synthesis
- GPU kernel and systems optimization
- Mathematical discovery such as circle packing, bijections, and quasi-Monte Carlo design
- Scientific computing and research workflows
- Quantum error-correction code discovery
- Improving LLM reasoning, mathematics, coding, and logical accuracy at inference time
- IBM Research used OpenEvolve for quantum error-correction code discovery.
- AMD-AGI Group used OpenEvolve in GEAK-OpenEvolve to optimize Triton GPU kernels for AMD Instinct GPUs.
- UC Berkeley's Sky Computing Lab used OpenEvolve for AI-Driven Research for Systems across systems-algorithm domains.