Skip to main content

Opt-AI Inc.

Category: AI Infrastructure

An automated on-device AI optimization platform that provides hardware-aware model compression, lightweighting, and inference acceleration for deploying complex AI onto resource-constrained edge devices. Opt-AI Inc. was founded in 2021. The company is led by Jaeho Lee. Based in Cheongju, South Korea. Team size: 10-50. Total funding raised: $3.5M+. Latest round: Strategic Investment (Undisclosed, Oct 2024). Key investors include FuturePlay, LG Electronics, Mashup Ventures, HL Mando.

Founded
2021
Headquarters
Cheongju, South Korea
Team size
10-50
Total funding
$3.5M+

Value proposition

Maximizes AI performance (inference speed and accuracy) while minimizing hardware resource consumption and deployment costs through a fully automated optimization pipeline.

Products and solutions

Opt-Hancer Platform (End-to-end automated AI optimization), Hardware-Aware NAS Engine (Neural Architecture Search), Model Compression & Quantization Toolkit (Specialized for ViT and sLM), NPU Orchestration & Inference Acceleration Solutions

Unique value

Unlike generic compression tools, Opt-AI provides 'Hardware-aware' optimization that redesigns neural architectures specifically for the target hardware's (NPU/GPU) unique architecture.

Target customer

Tier-1 automotive suppliers, defense contractors, robotics manufacturers, and consumer electronics companies requiring on-device AI capabilities.

Industries served

Automotive (In-vehicle monitoring, Autonomous driving), Defense & Aerospace (AI-based situational awareness), Robotics (Edge vision for service and industrial robots), Consumer Electronics (On-device sLM for mobile/home appliances)

Technology advantage

Offers a seamless, automated workflow from large model input to optimized deployment, significantly reducing the engineering time and expertise required for edge AI implementation.

How they differentiate

Differentiates through 'Hardware-Aware' optimization that automatically redesigns neural architectures based on specific NPU/GPU hardware constraints, rather than just pruning existing models.

Main competitors

Nota AI, SqueezeBits, Deci (Acquired by NVIDIA)

Key partnerships

LG Electronics (Superstart Lab, VS Division for in-vehicle camera platforms), HL Mando (Robotics camera platform optimization), Hanwha Systems (Defense-grade CPU-based AI optimization), Hyundai Motor Group (PoC for driver monitoring systems), FuturePlay, Mashup Ventures (Strategic Investors)

Notable customers

LG Electronics, HL Mando, Hanwha Systems, Hanwha Aerospace, Hyundai Motor Group (PoC)

Major milestones

Founded in Nov 2021 as a deep-tech spin-off., Secured initial seed funding from top-tier Korean VCs and HL Mando in 2023., Selected for 'Deep Tech TIPS' (1.5 billion KRW support) in July 2024., Received strategic investment from LG Electronics in late 2024., Won the CES 2025 Innovation Award for the Opt-Hancer platform.

Growth metrics

Selected for the prestigious 'Deep Tech TIPS' program in 2024; Received CES 2025 Innovation Award for 'Opt-Hancer'.

Market positioning

Deep-tech AI infrastructure provider specializing in edge/on-device optimization for automotive, defense, and robotics.

Geographic focus

South Korea (Primary), North America, and Europe (Expansion targeted via CES global showcases).

Patents and IP

Holds proprietary IP in Vision Transformer (ViT) quantization and Neural Architecture Search (NAS); recognized with high patent capability grades in industry databases (e.g., Innoforest).

About Jaeho Lee

Jaeho Lee is the Co-founder and CEO of Opt-AI Inc. He specializes in hardware-aware on-device AI optimization, specifically in Neural Architecture Search (NAS) and model compression. Under his leadership, Opt-AI has developed the 'Opt-Hancer' platform, which enables efficient deployment of AI models (including sLMs) on edge devices. He has secured strategic partnerships and investments from major corporations such as LG Electronics and Hyundai Motor Company. His background is rooted in deep-tech research at Chungbuk National University, leading to the company's recognition with the CES 2025 Innovation Award.

Official website: