DE
Search
  • Home
  • [Testseite Überarbeitung] Artificial Intelligence

[Testseite Überarbeitung] Artificial Intelligence

We provide Artificial Intelligence services designed to help clients leverage the power of AI to drive innovation and improvement. Our team can execute client projects from proof-of-concept to production, providing a range of services including data preparation, model development, testing, and deployment. We work closely with businesses to understand their needs and develop AI solutions that meet their specific requirements and goals. Our expertise in Generative AI includes Large Language Models (LLMs) and other Natural Language Processing methods with a focus on integration, training, hosting, and benchmarking.

Generative AI and Large Language Models

Generative AI

We've already developed and implemented several applications in the field of Generative AI. Some examples of applications include:

  • Al Assistant for Dataset Discovery via Text-to-SQL

  • Al Copilot for exposés in the real estate industry

  • Al Slide Deck generation

  • Atlassian Jira App Rock your Sprint Review

  • CaptAIn Bookworm (AI generated children's books)

  • Image/Portrait generation and editing

  • Multiple chatbots with and without data interactions

  • TAIA (TNG AI Assistant)

  • Text-to-3D Diffusion and Text-to-lmage Model Training

AI Productivity Tools

We also leverage custom AI productivity tools to boost efficiency for software engineering applications such as:

  • Automated unit test generation (Verlinkung der Landing Page folgt)

  • Automated end-to-end test generation (Verlinkung der Landing Page folgt)  

  • Automated code migration for Java and language conversions

  • Code and architecture analysis

AI Research

In the AI Research field, we are experimenting with various approaches and making progress in the following areas:

  • Model alignment and model surgery, e.g. DeepSeek-R1T2-Chimera

  • Design and implementation of agentic systems

  • Design, implementation, and evaluation of chatbots with and without data interaction (RAG)

  • Creation of a fine-tuned Optical Character Recognition (OCR) AI model based on olmOCR

  • Robotics: Utilizing teleoperation and reinforcement learning with a Unitree G1 humanoid robot, as well as co-development of the Robot Operating System version 2 and development of ROS/ROS2-applications

Methodical Expertise

All our work on Generative AI is supported by our strong methodical expertise such as:

  • Model and pipeline evaluation and benchmarking using advanced techniques such as LLM-based assessment

  • Domain-specific model fine-tuning for e.g. legal tech, source code, and German language

Computer Vision

Our team has expertise in computer vision solutions for image analysis, object detection, and automated quality control across various industries. We can help to develop applications that can interpret and understand visual data, enabling use cases such as quality control, security, and autonomous systems.

  • Realtime Deepfakes

  • Defect detection and quality control in different industries, e.g. agriculture, high-tech manufacturing, and pharmaceutical wholesale

  • Autonomous driving and flight, including Simultaneous Localization and Mapping (SLAM) as well as controlling

  • Motion analysis for sports athletes

  • Biometric identification, including face identification, iris recognition, and Automated Fingerprint ldentification System (AFIS)

  • Security applications for public spaces

    • Video anonymization

    • Touchless gesture feedback

    • Unattended luggage detection

Classical Machine Learning

We develop and implement also classical machine learning solutions for industrial applications, enabling businesses to optimize processes, improve efficiency, and reduce costs. Some project examples are the following:

  • Fraud and risk detection in fintech and telecommunications industry

  • Predictive costing in global supply chains

  • Size recommender for a webshop

Reinforcement Learning

Our expertise in reinforcement learning empowers us to develop cutting-edge AI solutions tailored to business needs. Projects in the area of industrial process optimization included:

  • Optimizing Maintenance Decisions with Monte Carlo tree search (MCT)

  • Satisfiability Problem (SAT) Solving

  • Automatic adjustment of the steering parameters for an industrial extrusion machine

Platforms & Hosting

We host our AI models on own GPU clusters to maintain control over sensitive data. Our portfolio for clients includes support for on-premises solutions, cloud-services, and edge devices:

On-Premises

  • Hosting of LLMs and Al Services on a GPU Cluster, including embedding, transcription, translation, and re-ranking

  • Fine-tuning and training of LLMs

  • Infrastructure administration

Cloud-Services

  • AWS ML Services

  • Azure Al Services and Azure OpenAI

  • Google Vertex Al

  • Lambda Cloud

Edge Devices

  • ESP32

  • Raspberry Pi with Neural Compute Stick

  • Nvidia Jetson Nano

Interested?

Contact us via info@tngtech.com or use one of the other contact options to learn more about our AI services. We are happy to support you.