LG Aimers 8th: Taking on the Challenge with ‘EXAONE,’ Korea’s Representative AI
“Technology and innovation that change the world begin with people.” 
Image 1. LG Aimers 8th Participants
Today, Artificial Intelligence (AI) has established itself as a core technology across various industries, proving its vast potential. However, to apply AI to real-world industrial sites, it is essential to have not only intelligent models but also practical optimization and compression capabilities that allow these models to run efficiently on individual devices.
From the 4th, LG AI Research hosted the ‘LG Aimers Hackathon,’ a hands-on program designed to solve industrial challenges using AI technology, for two days and one night at the LG Inhwawon in Icheon, Gyeonggi-do. Now in its 8th, the competition was held as part of a youth talent education initiative driven by LG’s philosophy that “Technology and innovation that change the world begin with people.”
LG Aimers is South Korea's largest AI education program, with over 20,000 young participants since its launch in the second half of 2022 through the first half of this year. For this 8th LG Aimers offline hackathon, 94 young talents across 27 teams participated, having been selected from a total of 2,339 applicants after completing online training and preliminary rounds.
Key Task: Compressing the EXAONE Model
Image 2. LG Aimers 8th Offline Hackathon Site
The theme for this hackathon, presented by LG AI Research, was ‘Developing a Compressed EXAONE Model.’
‘Large scale Language Model (LLM) Compression’ is an optimization technology that reduces the size and improves the inference speed of AI models while maintaining their performance and accuracy. This is a critical technology for implementing 'On-device AI,' which allows AI models to operate independently inside individual devices such as laptops or smartphones. The participating students took on the challenge of advancing 'EXAONE,' the representative AI of Korea, and successfully developed a total of 27 compressed models.
Reflecting on the experience, one student said, "Working on the EXAONE 4.0 1.2B model was a priceless opportunity to apply the theoretical knowledge I learned in the lab to an actual model and see the results firsthand." Mentors from the EXAONE Lab, who spent the intensive two days and one night alongside the students, also offered words of support. "Producing tangible results under such high pressure in a short time is a massive achievement in itself," he emphasized. "I hope participants use this success as a stepping stone to build even greater confidence in your future endeavors."
Field-Oriented Career Support Activities
Image 3. LG Aimers 8th Job Fair
On the second day of the competition, a job fair was held with the participation of eight major LG affiliates, including LG AI Research, LG Electronics, LG Display, LG Chem, LG Energy Solution, LG Uplus, and LG CNS. HR managers from each company shared recruitment information, provided practical career counseling, and guided participants through registration for the ‘LG AI Talent Pool,’ which offers priority consideration for future hiring.
In addition, active support was provided to enhance the employment competitiveness of these young talents through 1:1 customized consulting and mock AI interview lectures, ensuring they are equipped with both practical skills and job readiness.
The Challenge for High-Performance Compressed Models
Image 4. LG Aimers 8th Hackathon Final Winning Teams Presentation
(Left) 1st place: Hand Cream Mice team, (Middle) 2nd place: Miku-chan team, (Right) 3rd place: Keroro Platoon team
As the intense two-day hackathon drew to a close, the final winning teams were selected for demonstrating exceptional technical prowess in compressing and optimizing the models. The awards ceremony, which highlighted both technical maturity and a spirit of challenge, saw three teams taking home the honors. The Grand Prize, the Minister of Employment and Labor Award, was awarded to team ‘Hand Cream Mice’. By leveraging W8A8 quantization, ‘Hand Cream Mice’ team topped both the leaderboard and the overall rankings with a brilliant strategy that maximized memory efficiency with minimal loss in performance. The second place, the LG AI Research Award, went to team ‘Miku-chan’, who received high marks for their experimental spirit in implementing DPO (Direct Preference Optimization) and exploring a diverse range of methodologies. The third place was taken by team ‘Keroro Platoon’, who demonstrated stable technical execution and robust model optimization skills.
The Grand Prize winners, Team ‘Hand Cream Mice,’ shared their excitement after the ceremony. "We were able to maximize performance because we never gave up and constantly exchanged ideas within the team," they said with a smile. "We are so happy to achieve this valuable result together after all the hard work we put in."
Image 5. LG Aimers 8th Hackathon Final Winning Teams Awards Ceremony
A Continuing Journey with Young AI Talent
LG Aimers offers the entire process—from AI graduate school-level lectures to the hackathon and career support—completely free of charge. Anyone aged 19 to 29 with basic AI knowledge and coding skills can apply. The program runs twice a year (during summer and winter vacations) in consideration of academic schedules.
Recruitment for the next term, LG Aimers 9th, is scheduled to begin this coming May. This will be an excellent opportunity for aspiring AI experts to handle real-world industrial data and receive mentoring from professionals.