CogiTechCogitechSoftware Development Lab

Our team delivers advanced software solutions tailored for complex problems, focusing on efficiency and innovation.

Enterprise
Technical Specification Assistant

Technical Specification AssistantVisitVideo

With Tessa, the process of creating technical specifications becomes easier and faster. You can define goals and requirements through simple dialogue with our AI model, customized to provide exactly what the specifications need. The prompting is designed to be natural and still encourage the user to create complete specification documents, filling out every important section that such a document requires.

Highlights

  • Fine-Tuned models for technical specification design
  • Responsive prompting with real-time document updates
  • User prompt classification and reasoning with agentic workflows
  • High-volume WebSocket serverless architecture
Automate and optimise material cutting

Automate and optimise material cuttingVisit

This project aims to automate and optimize the cutting, milling, and storage of materials. By automating the cut planning process, our solution minimizes material remains and reduces waste, driving greater productivity and sustainability. Supported by the European Institute of Technology (EIT) initiative, we leverage cutting-edge software techniques to enhance efficiency and precision in manufacturing operations.

Highlights

  • Automated ahead-of-time order fulfillment planning
  • Digitization of inventory and electronic warehouse tracking
  • AI inventory purchasing prediction and stock optimization
Find and redact PII in documents

Find and redact PII in documentsVisit

Redacted.ai is a document detection and redaction suite. It is able to detect and redact protected information in a wide variety of document and image formats. The AI models the system employes are all designed to work as part of a cloud solution, and it is available on all major cloud providers, as well as on-prem. Redacted.ai was acquired by OneTrust as part of its DSAR offering.

Highlights

  • Advanced OCR and computer vision for documents and images
  • Detection of different types of information (medical, personal, facial recognition, ...)
  • Enterprise architecture and DevOps for AI models and document processing
  • Wide document type compatibility (PDF, DOCX, XML, MSG, ...)
  • Multi-Cloud deployment includng AWS, Azure, and GCP
Get context, delve deeper, and learn more

Get context, delve deeper, and learn moreVisitVideo

The goal of InFold is to simplify news tracking for readers. We crawl the web with AI workers which are designed to pick up a variety of data, such as text, image and video and use natural langauge processing to classify all of it into distinct topics. This topic modeling is enriched with relevant sources, such as social media posts and relevant video content. We present this content to the user in an information-rich, though ordered and easily understood environment, continuously improved to take their preferences into account.

Highlights

  • Wide-spectrum data aggregation from across the internet
  • Natural language analysis and clustering for topic modeling
  • LLM chatbots for news topics and user interaction
  • Asynchronous queue processing for thousands of articles and keywords per second
Collectively draw on a shared canvas

Collectively draw on a shared canvasVisit

A social game where players collaborate on a shared canvas by changing the color of individual pixels. The goal is to illustrate the process of self-organization within a community of random players with the help of blockchain technology.

Highlights

  • Social gaming incentives and reward mechanisms
  • Multiple token standards (ERC20, ERC721, and ERC1155)
  • Blockchain data scraping and extraction with subgraphs