Delta AI — Pioneering Intelligent Solutions

What We Do

End-to-End Intelligent Solutions

From prototype to publication — we cover every stage of your AI journey with uncompromising expertise.

🤖

AI & Machine Learning

Custom model development, fine-tuning, and deployment — LLMs, Vision Transformers, GANs, Diffusion Models, and beyond.

PyTorchTensorFlowHuggingFaceONNX

👁

Computer Vision & AR/VR

Real-time perception systems — hand tracking, depth estimation, activity recognition, spatial computing, and XR interfaces.

MediaPipeOpenCVDepthAnythingUnity

🔐

Network & Cybersecurity

Security architecture, penetration testing advisory, intrusion detection via ML, and enterprise network design.

Network DesignIDS/IPSThreat Modeling

📝

Research & Publication

End-to-end academic guidance — experiment design, data collection, paper writing, and peer review prep for top-tier venues.

CVPRACLNeurIPSICCVICDE

🎓

University Admissions Consulting

Strategic guidance for graduate and undergraduate admissions globally — SOP, LOR, research portfolio, and interview coaching.

South KoreaUSAEuropeWorldwide

⚙

Embedded & Edge AI

Deploying AI on resource-constrained hardware — FPGA, microcontrollers, Raspberry Pi, Jetson. Signal processing and MATLAB.

MATLABJetsonFPGATensorRT

Why Delta AI

Where Academia Meets
Industrial Precision

Delta AI is founded by active researchers who publish at top venues and simultaneously build production systems. You get both worlds — rigorous methodology and pragmatic delivery.

The Δ (Delta) symbol is mathematics for change and difference — exactly what we bring to your challenges.

✅

Peer-Reviewed Excellence Our methods are validated at CVPR, ACL, EMNLP, ICDE — not just sales decks.

🌏

Global, Remote-First Delivery Headquartered in South Korea, serving clients worldwide across time zones.

⚡

Full-Stack AI Capability From raw sensor data to deployed model to published paper — we own the full pipeline.

h-index (Google Scholar)

SCIE Journal Papers

55+

GitHub Repositories

193+

Total Academic Citations

The Team

Founded by Researchers,
Built for Results

Deep expertise, proven track record, and a shared passion for making AI work in the real world.

Founder & CEO

Dr. Mukhiddin Toshpulatov

PhD · Asst. Professor, Gachon University & Research Professor, Inha University

Leading researcher in Computer Vision, AI-driven HCI, Sign Language Technologies, and Generative AI. Author of publications at CVPR, ACL, EMNLP, ICDE, and multiple high-impact journals. 25+ years in mathematics, computer science, and embedded systems across Uzbekistan, South Korea.

Computer Vision Generative AI AR/VR/XR NLP Medical AI Embedded Systems

🌐 Portfolio 💻 GitHub ✉ Email

Co-Founder & COO

Chingiz Tursunbaev

PhD Candidate at BMSE & Researcher at XVoice AI Lab, Inha University

As a Ph.D. candidate at Inha University, Chingiz bridges the gap between complex academic research and real-world AI applications. He specializes in designing deep learning architectures for video processing and accessibility tech. Recent Academic Contributions: ACL/EMNLP, NeurIPS, Bigcomp, KCC/KSC.

Computer Vision Foundation Models (SAM 2, SAM 3) Deep Learning Sign Language Recognition (SLR) Systems GPU Server Administration Linux Workstation Management

🌐 Portfolio 💻 GitHub ✉ Email

Co-Founder & CTO

Nirmal Adhikari

PhD Candidate at BMSE & Researcher at XVoice AI Lab, Inha University

Working on machine learning, computer vision, multimodal learning, and intelligent systems. Research focuses on understanding, developing, and evaluating AI models for real-world applications, with experience in sign language recognition, sequence modeling, representation learning, and deep learning-based perception systems.

Computer Vision Generative AI Vision LLM NLP, LLM Medical AI Foudational Models

🌐 Portfolio 💻 GitHub ✉ Email

Co-Founder & CKO

Giyosbek Rustamov

Integrated MS & PhD Candidate at BMSE & Researcher at XVoice AI Lab, Inha University

Giyosbek focuses on advancing multimodal AI systems that integrate speech and visual information. His research interests include Speech Emotion Recognition, Visual Speech Recognition (lip reading), and Agentic AI systems. He is particularly interested in developing intelligent systems that can understand, interpret, and interact with human communication in real-world environments. Academic Contributions: NeurIPS, KCC

Visual Speech Recognition Generative AI Vision LLM Speech Emotion Recognition Medical AI Agentic AI

🌐 Portfolio 💻 GitHub ✉ Email

Co-Founder & CIO

Sherali Juraliev

Software Engineer & Research assistant at XVoice AI Lab, Inha University

2nd-year undergraduate in Integrated Systems Engineering at Inha University. Currently involved in active CVPR 2026 research on real-time fingertip contact detection via depth and motion fusion.

Computer Vision Depth Models Vision LLM Software Engineering Medical AI Backend Development

🌐 Portfolio 💻 GitHub ✉ Email

Co-Founder & CMO

Ulugbek Amankulov

Software Developer · Business Technology Strategist · Video Content Creator · Web Designer

Multilingual technologist (Korean TOPIK 5, English TEPS) blending software development with business acumen. Certified in GTQ Hangul & Photoshop. Bridges technical delivery with market strategy and brings fresh innovation to every project.

Software Development Business Strategy Korean / English Design (GTQ) Image, Video Editor

🌐 Portfolio 💻 GitHub ✉ Contact

Published Research

Validated at the World's
Top Venues

Our consulting is grounded in research we actually publish — not just theory.

CVPR 2026

Real-Time Multimodal Fingertip Contact Detection via Depth and Motion Fusion for Vision-Based HCI

M. Toshpulatov, W. Lee, S. Lee, G. Lee — IEEE/CVF Conference on Computer Vision and Pattern Recognition 2026

SCIE Journal

Human Pose, Hand and Mesh Estimation Using Deep Learning: A Survey

M. Toshpulatov, W. Lee, S. Lee, A. Haghighian Roudsari — The Journal of Supercomputing 78(6), 7616–7654 · 72 citations

SCIE Journal

Talking Human Face Generation: A Survey

M. Toshpulatov, W. Lee, S. Lee — Expert Systems with Applications 219, 119678 · 48 citations

SCIE Journal

Generative Adversarial Networks and Their Application to 3D Face Generation: A Survey

M. Toshpulatov, W. Lee, S. Lee — Image and Vision Computing · 65 citations

SCIE Journal

Deep Learning Pathways for Automatic Sign Language Processing

M. Toshpulatov, W. Lee, J. Jun, S. Lee — Pattern Recognition, 111475 · 18 citations

SCIE Journal

DDC3N: Doppler-Driven Convolutional 3D Network for Human Action Recognition

M. Toshpulatov, W. Lee, S. Lee, H. Yoon, U. Kang — IEEE Access · 17 citations

SCIE Journal

Hyperspectral Anomaly Detection with Enhanced Spectral Graph Transformer Network

F. Safarov, M. Toshpulatov et al. — IEEE Access 13, 170554–170576 · 2025 · 3 citations

SCIE Journal

Adaptive Feature Refinement for Texture-Preserving Single Image Super-Resolution

M. Toshpulatov, F. Safarov et al. — Cluster Computing 29(3), 152 · 2026

IEEE Conference

Enhancing Sign Language Recognition: A Computationally Efficient and Interpretable Multi-Modal Approach

C. Tursunbaev, M. Toshpulatov et al. — IEEE BigComp 2026

IEEE Conference

Privacy-Preserving of Human Identification in CCTV Data Using Deep Learning

M. Toshpulatov, A. Haghighian Roudsari et al. — IEEE BigComp 2022 · 5 citations

IEEE Conference

Human Action Recognition Utilizing Doppler-Enhanced Convolutional 3D Networks

M. Toshpulatov, W. Lee — IEEE BigComp 2024

IEEE Conference

Research Issues on Generative Adversarial Networks and Applications

M. Toshpulatov, W. Lee, S. Lee, R. Tojiboev — IEEE BigComp 2020 · 17 citations

View Full Google Scholar Profile →

Intelligence That
Changes Everything
Δ = Difference. AI = Future.

End-to-End Intelligent Solutions

AI & Machine Learning

Computer Vision & AR/VR

Network & Cybersecurity

Research & Publication

University Admissions Consulting

Embedded & Edge AI

Where Academia Meets
Industrial Precision

Founded by Researchers,
Built for Results

Dr. Mukhiddin Toshpulatov

Chingiz Tursunbaev

Nirmal Adhikari

Giyosbek Rustamov

Sherali Juraliev

Ulugbek Amankulov

Validated at the World's
Top Venues

Real-Time Multimodal Fingertip Contact Detection via Depth and Motion Fusion for Vision-Based HCI

Human Pose, Hand and Mesh Estimation Using Deep Learning: A Survey

Talking Human Face Generation: A Survey

Generative Adversarial Networks and Their Application to 3D Face Generation: A Survey

Deep Learning Pathways for Automatic Sign Language Processing

DDC3N: Doppler-Driven Convolutional 3D Network for Human Action Recognition

Hyperspectral Anomaly Detection with Enhanced Spectral Graph Transformer Network

Adaptive Feature Refinement for Texture-Preserving Single Image Super-Resolution

Enhancing Sign Language Recognition: A Computationally Efficient and Interpretable Multi-Modal Approach

Privacy-Preserving of Human Identification in CCTV Data Using Deep Learning

Human Action Recognition Utilizing Doppler-Enhanced Convolutional 3D Networks

Research Issues on Generative Adversarial Networks and Applications

From First Call to Final Delivery

Discovery Call

Proposal & Scope

Execution

Delivery & Support

Ready to Build Something
Extraordinary?

Let's Work
Together

Intelligence That Changes Everything Δ = Difference. AI = Future.

End-to-End Intelligent Solutions

AI & Machine Learning

Computer Vision & AR/VR

Network & Cybersecurity

Research & Publication

University Admissions Consulting

Embedded & Edge AI

Where Academia MeetsIndustrial Precision

Founded by Researchers,Built for Results

Dr. Mukhiddin Toshpulatov

Chingiz Tursunbaev

Nirmal Adhikari

Giyosbek Rustamov

Sherali Juraliev

Ulugbek Amankulov

Validated at the World'sTop Venues

Real-Time Multimodal Fingertip Contact Detection via Depth and Motion Fusion for Vision-Based HCI

Human Pose, Hand and Mesh Estimation Using Deep Learning: A Survey

Talking Human Face Generation: A Survey

Generative Adversarial Networks and Their Application to 3D Face Generation: A Survey

Deep Learning Pathways for Automatic Sign Language Processing

DDC3N: Doppler-Driven Convolutional 3D Network for Human Action Recognition

Hyperspectral Anomaly Detection with Enhanced Spectral Graph Transformer Network

Adaptive Feature Refinement for Texture-Preserving Single Image Super-Resolution

Enhancing Sign Language Recognition: A Computationally Efficient and Interpretable Multi-Modal Approach

Privacy-Preserving of Human Identification in CCTV Data Using Deep Learning

Human Action Recognition Utilizing Doppler-Enhanced Convolutional 3D Networks

Research Issues on Generative Adversarial Networks and Applications

From First Call to Final Delivery

Discovery Call

Proposal & Scope

Execution

Delivery & Support

Ready to Build SomethingExtraordinary?

Let's WorkTogether

Intelligence That
Changes Everything
Δ = Difference. AI = Future.

Where Academia Meets
Industrial Precision

Founded by Researchers,
Built for Results

Validated at the World's
Top Venues

Ready to Build Something
Extraordinary?

Let's Work
Together