Domyn builds Colosseum 355B, a sovereign AI foundation model, using NVIDIA DGX Cloud
Domyn · Italy
Domyn (formerly iGenius), an Italian AI company serving highly regulated sectors such as financial services and public administration, used NVIDIA DGX Cloud with over 3,000 NVIDIA H100 GPUs to continue-pretrain Colosseum 355B, a 355-billion-parameter foundation LLM. Within one week Domyn had access to the dedicated infrastructure, and within two months completed continued pretraining, achieving 82.04% accuracy on the MMLU benchmark. The model powers Domyn's business intelligence agent, Crystal, a sovereign AI solution deployed on private infrastructure.
Overview
Domyn (formerly iGenius), an Italian AI company serving highly regulated sectors such as financial services and public administration, used NVIDIA DGX Cloud with over 3,000 NVIDIA H100 GPUs to continue-pretrain Colosseum 355B, a 355-billion-parameter foundation LLM. Within one week Domyn had access to the dedicated infrastructure, and within two months completed continued pretraining, achieving 82.04% accuracy on the MMLU benchmark. The model powers Domyn's business intelligence agent, Crystal, a sovereign AI solution deployed on private infrastructure.
This entry has 12 published fields tied to exact passages in an immutable source capture.
Inspect the highlighted sourceThe challenge
Domyn aimed to develop a state-of-the-art foundational LLM within a tight timeline but faced challenges accessing large-scale GPU clusters (thousands of GPUs) and securing support for highly scalable training frameworks.
The solution
Domyn collaborated with NVIDIA to build Colosseum 355B using NVIDIA DGX Cloud infrastructure with over 3,000 H100 GPUs on a dedicated, high-bandwidth RDMA-based network with 500 TB of Lustre-based high-performance storage and the NVIDIA NeMo Framework. Domyn used continued pretraining (CPT) in FP8 precision via NeMo Framework's Transformer Engine, followed by supervised fine-tuning and Direct Preference Optimization (DPO) using NVIDIA NeMo Aligner. The model powers Domyn's business intelligence agent, Crystal, a sovereign AI solution built as an end-to-end stack including database integration, AI-assisted configuration, LLM-powered orchestration for tool usage, query execution and generation, and private deployment infrastructure.
Reported business value
In less than one week, Domyn had access to dedicated large-scale infrastructure with over 3K GPUs, and within two months had completed continued pretraining for Colosseum 355B, achieving 82.04% accuracy on the MMLU benchmark in a 5-shot setting. By optimizing training configuration, Domyn improved Model FLOP/s Utilization (MFU) from an initial 25% to 40% in BF16, and accelerated the overall training step 1.15x by moving to FP8.
Sources
Open any source and check the claim yourself — that is the point of the register.
Other technology & software entries in the register.
Supermetrics: Helping Marketers Redefine Efficiency with AI-Powered Data Analysis
Supermetrics, a Finland-based marketing intelligence platform serving 15,000+ customers across 132 countries, built an AI agent on Google Cloud using Vertex AI Agent Builder and the Agent Development Kit (ADK) that autonomously manages data connections, fixes pipeline errors, and analyzes campaign performance in real time, suggesting new creative options using Imagen. The agent automates the weekly marketing reporting cycle that previously took performance marketers up to four hours, reclaiming over 15 hours per month per marketer for strategy and creative testing. The system uses a central AI agent that interprets natural language requests and delegates tasks to sub-agents, and stores 'core memories' of user preferences for personalized context.
Strava's Athlete Intelligence Translates Workout Data into Simple and Personalized Insights
Strava launched Athlete Intelligence, an AI-powered feature available as a public beta to subscribers, which analyzes and interprets workout data across pace, heart rate, elevation, power, and Relative Effort into simple, personalized insights and guidance. The feature spots 30-day performance trends, detects milestones such as fastest pace or longest distance, and offers tailored feedback for each activity, drawing on more than 10 billion activity uploads on Strava.
Fifth Dimension unlocks insights and intelligence with Google Cloud
Fifth Dimension, founded in 2023, provides an AI platform combining machine learning, predictive analytics and natural language processing to uncover hidden patterns in real estate data, processing over 1TB of data per month by 2025. Facing scalability and cost problems with its original infrastructure, the company adopted Google Cloud, building its ML stack on Vertex AI (running Google Cloud Gemini and Anthropic Claude models), plus Cloud SQL, Cloud Run and Pub/Sub. This scaled processing capacity 50x to handle document-processing surges, supported 6x global client growth across multiple regions, and decreased infrastructure costs by 30% through serverless architecture. Model deployment time dropped from weeks to days, platform engagement tripled in 2025, and annual recurring revenue grew 6x in the same year.
Spotlight: Qodo Innovates Efficient Code Search with NVIDIA DGX
Qodo, a multi-agent code integrity platform, built its AI agents on retrieval-augmented generation powered by a state-of-the-art code embedding model trained on NVIDIA DGX. Qodo fine-tuned two embedding models, Qodo-Embed-1-1.5B and Qodo-Embed-1-7B (based on Qwen), achieving state-of-the-art accuracy on the Hugging Face MTEB CoIR leaderboard in their size categories. In a collaboration with NVIDIA, Qodo's code indexer, RAG retriever, and embedding model were substituted into NVIDIA's internal RAG solution (Genie) for searching private code repositories, integrated into NVIDIA's internal Slack system, yielding more detailed and accurate responses to technical questions from expert C++ developers than the original pipeline, evaluated using Ragas-generated synthetic questions against RTXDI, RTXGI and RTXPT SDK repositories.
Was this helpful?
Your feedback helps us improve our use case database