[AI Revolution](216) Team Sparta: "Claude Sonnet-Level Coding in Closed Networks... Unleashing Corporate Potential"

From Education to Technical Infrastructure… Supporting AI Upskilling
Securing Success with Closed-Network AI That Reduces Data Leakage Concerns
Advancing Toward an Industry-Specific AI Agent Platform

"We aim to help people working in closed networks utilize artificial intelligence (AI) by developing a model that can deliver Claude Sonnet-level performance even on a palm-sized graphics processing unit (GPU)."

Beomgyu Lee, CEO of Teamsparta. Teamsparta.

Beomgyu Lee, CEO of Team Sparta, an AI upskilling company, recently shared with The Asia Business Daily his vision to support corporate AI transformation (AX) with AI that operates on internal networks disconnected from the outside internet.

Team Sparta, which got its start as a coding education company, has transformed into an 'AI upskilling' business dedicated to enhancing AI competency for both individuals and organizations. Following the launch of its corporate AI education and AX solutions business in 2024, Team Sparta began developing AI solutions tailored for closed networks at the end of last year. The ambition is to move beyond simply teaching employees how to use AI—Team Sparta wants to create an environment where acquired skills can be applied directly to real work. Lee explained, "Team Sparta’s strength lies in its ability to build AI competency among people as well as provide the technical infrastructure to support it."

This direction stems from Lee's experience with Woowa Brothers and BonAngels Venture Partners. Observing the process of solving problems with technology and turning ideas into startups sparked his interest in reducing the barriers to entry for technology. Lee said, "I realized that beyond creating good technology, it is just as important to make it easy for more people to learn and use technology to solve their own problems," highlighting the background to his entrepreneurial journey.

"Bringing the 'ChatGPT Moment' to Closed-Network Enterprises"

Team Sparta is targeting sectors where security and data management are critical, such as public, healthcare, and finance. Because these industries handle sensitive data, it has been difficult to use external AI solutions, while building large-scale models internally imposes a heavy infrastructure burden. The creation of Team Sparta's proprietary foundation model, 'K-AX Spartan,' was aimed at providing technology that fits these types of corporate environments.

The strength of K-AX Spartan is that it delivers the required performance for businesses even with limited computing resources. This ultra-lightweight model features 1.8 billion parameters and ranked first in the high-level inference evaluation (HLE) of the K-AI Leaderboard, a domestic AI performance evaluation platform. Compared to the average size of the top 10 models in the evaluation, it achieved high performance while being roughly 20 times smaller. Lee stressed, "Rather than trying to compete in the same way as massive general-purpose models, our focus is on developing industry-optimized models and making them practically usable within internal networks."

Team Sparta is continuing to advance 'AX Fortress,' an AI solution tailored for closed networks. Serving as a coding-specialized agent that assists work without transferring data externally, it is currently being expanded to areas such as manufacturing data analysis and security operations log analysis. Lee remarked, "For customers in closed networks who have so far been unable to use AI, this could be a transformative experience akin to encountering ChatGPT for the first time."

In practice, AX Fortress is currently undergoing proofs-of-concept (PoC) and adoption testing with companies in manufacturing, finance, and security. Last month, Team Sparta was selected as one of the top 20 teams in the Asia-Pacific region at the NVIDIA Inception Grand Challenge, a prestigious AI startup competition. At the finals held on the 23rd, they showcased a solution built on AX Fortress that addresses supply chain issues for manufacturing companies.

From Manufacturing to Finance... "Becoming the Top Corporate AI Solution"

Team Sparta's future goal is to develop AX Fortress into an enterprise AI agent platform that can perform specialized work in various industries. The company plans to add specialized agents for each business function atop a common platform for safe internal data usage, enabling expansion into areas such as supply chain management for manufacturers and document review for financial companies.

This year, Team Sparta aims to secure flagship case studies that demonstrate AX Fortress’s capabilities at client sites and focus on upgrading the product so that it can be applied repeatedly across multiple sectors. Lee stated, "We want to become the go-to enterprise AI solution for businesses considering AI adoption due to concerns over security or data," adding, "We hope to help each organization take the next step with AI tailored to their needs and to become the driving force that unlocks its potential."

This content was produced with the assistance of AI translation services.