Protech
Dienstleistungen
2 Dienstleistungen angeboten von Protech
Portfolio
7 Projekte durchgeführt von Protech
Projekt für Confidential in der Hardwaretechnologie & Equipment Branche für ein B2B Publikum.
<p><strong>Short description</strong></p><p>Analyzing blood smears manually under a microscope is slow and labor-intensive. Protech designed, developed, and deployed a fully automated, cloud-based web solution that detects, counts, and analyzes blood cells in microscopic blood smear images. Hosted entirely on AWS, the system ingests high-resolution images from electron microscopes, runs them through a computer vision and machine learning pipeline for digital staining and blood cell classification, and stores the results. Each stage of the pipeline runs as an independent microservice on AWS, allowing the workflow to scale reliably. The entire workflow, from image ingestion to final results, runs with a single click and no manual intervention, with detailed processing logs shown on completion.</p><p></p><p><strong>Use case</strong></p><p>"As an electron microscope manufacturer, I want high-resolution blood smear images to upload automatically to the cloud and trigger a pipeline that delivers results for analysis by microbiologists through a web interface."</p><p></p><p><strong>Result</strong></p><p>Replaced manual microscope analysis with a one-click, fully automated cloud pipeline requiring no operator intervention.</p><p></p><p><strong>Expertise</strong></p><p>Computer vision</p><ul><li>Deep learning</li><li>Automated image-processing pipeline</li><li>Digital staining and cell classification</li><li>Frontend and backend development</li><li>Microservices architecture</li><li>AWS cloud deployment</li></ul>Projekt für Confidential in der Software & Computerservice Branche für ein B2B Publikum.
<p><strong>Short description</strong></p><p>Manual car damage assessment for insurers, rental, and car-sharing companies is slow and inconsistent. Protech designed and developed a full pipeline of algorithms that automate it: detection and classification of damage, camera pose estimation, car make/model/color recognition, a 2D/3D registration algorithm that maps damage from 2D images onto a 3D model of the vehicle, and damage measurement modules. Protech also optimized the algorithms and built the cloud serving infrastructure.</p><p></p><p><strong>Use case</strong></p><p>"As a manager at an insurtech company, I want to speed up and automate car damage evaluation for my clients and partners by putting state-of-the-art AI into the process."</p><p><strong></strong></p><p><strong>Result</strong></p><p>Automated the full damage-assessment pipeline, from raw images to mapped, measured damage on a 3D model of the vehicle.</p><p></p><p><strong>Expertise</strong></p><ul><li>Computer vision</li><li>Machine learning and deep learning</li><li>Defect detection and classification</li><li>Vehicle make, model, and color recognition</li><li>Camera pose estimation</li><li>2D/3D registration</li><li>Damage measurement</li><li>Algorithm optimization</li><li>Cloud serving infrastructure</li></ul>Projekt für Confidential in der Software & Computerservice Branche für ein B2B Publikum.
<p><strong>Short description</strong></p><p>Onboarding new customers for IT and telecom providers traditionally required a branch visit or slow, error-prone manual data entry, and left room for identity fraud. Protech developed a mobile application that automates the entire process, letting new or existing customers open and manage a contract remotely from their own phone. Using only the device's built-in camera, the app detects and reads identity documents, extracts the required data, verifies the customer through face matching, and supports digital signing of the contract. The flow is designed to complete with minimal user interaction and no assistance from the provider's staff, while actively blocking attempts at identity fraud.</p><p></p><p><strong>Use case</strong></p><p>"As a mobile operator, I want customers to buy a product and sign a contract from home with no help from my sales agents, so I can cut onboarding workload while still capturing accurate identity data and preventing identity fraud."</p><p><strong></strong></p><p><strong>Result</strong></p><p>Enabled fully remote, self-service onboarding, removing the need for branch visits and manual data entry.</p><p></p><p><strong>Expertise</strong></p><ul><li>Computer vision</li><li>Machine learning and deep learning</li><li>Identity document detection and classification (SVM, CNN)</li><li>OCR and text extraction</li><li>QR and MRZ code detection and extraction</li><li>Face detection and matching (deep CNN)</li><li>Digital document signing</li><li>iOS and Android application development</li><li>On-device CNN optimization for mobile</li></ul>Projekt für Confidential in der Software & Computerservice Branche für ein B2B Publikum.
<p><strong>Short description</strong></p><p>Shopnosis is a shopper-research company that studies how customers behave in stores and delivers those insights to its retail clients. It conducts in-store recordings in which respondents wear camera glasses while shopping, and Protech developed the analytics platform that turns this footage into behavioral insights: shopper footpath, which products and in-store marketing communications (marcoms) they see, what they touch, hold, buy, or return, time spent looking at items, hit ratio, observation sequence, and seen-to-buy ratio. The platform is delivered as a web application built on a microservices architecture on Azure, with a frontend for reviewing results and backend services that process the captured footage. This lets Shopnosis run large-scale, cost-effective quantitative studies with hundreds or thousands of shoppers and deliver the results to retailers.</p><p></p><p><strong>Use case</strong></p><p>“As a shopper-research company serving retailers, I want to record real in-store shopping and have a platform automatically analyze the footage at scale, so I can deliver reliable behavioral insights to my retail clients faster and more cost-effectively than manual studies.”</p><p></p><p><strong>Result</strong></p><p>Turned expensive small-sample studies into large-scale quantitative analysis of real shopper behavior.</p><p></p><p><strong>Expertise</strong></p><ul><li>Computer vision</li><li>Deep learning</li><li>Data analytics</li><li>Frontend and backend development</li><li>Microservices architecture deployed on Azure</li></ul>Projekt für Confidential in der Unterhaltung & Events Branche für ein B2B / B2C Publikum.
<p><strong>Short description</strong></p><p>Users new to fitness and nutrition rarely have affordable, personalized expert guidance. Protech designed, developed, and deployed an intelligent fitness coach built on large language models. Inside a mobile app, users chat with the AI Coach about fitness, nutrition, and wellbeing. The coach combines multiple LLMs (Claude, Mistral, Amazon Titan) on AWS Bedrock with a retrieval-augmented knowledge base, so answers reflect each user's profile and goals. On request it generates personalized training or nutrition plans drawing on an existing library of exercise videos and recipes. The solution runs on a serverless microservices architecture on AWS for scalability.</p><p></p><p><strong>Use case</strong></p><p>"As a user with little fitness or nutrition experience, I want an AI Coach in a mobile app that answers my questions and builds a training and meal plan tailored to my goals and health profile, with helpful videos and recipes in one place."</p><p></p><p><strong>Result</strong></p><p>Delivered a scalable, personalized coaching experience that gives non-expert users tailored guidance on demand.</p><p></p><p><strong>Expertise</strong></p><ul><li>NLP and generative AI</li><li>Retrieval-augmented generation (RAG)</li><li>Multi-LLM integration (Claude, Mistral, Amazon Titan)</li><li>Serverless microservices architecture</li><li>Frontend and backend development</li><li>Mobile application development</li></ul>Projekt für Confidential in der Hardwaretechnologie & Equipment Branche für ein B2B Publikum.
<p><strong>Short description</strong></p><p>Automated inspection equipment for agri-food must judge the quality of every kernel to remove defective product. Protech developed a vision-based algorithm for automated quality inspection and sorting: it detects and classifies damage on peanut kernels, sorts them by configurable damage thresholds, and measures size (major dimensions and volume estimation) while distinguishing whole from split kernels across all peanut types. The algorithm runs at production-line speed, assessing and sorting each kernel in real time as it passes through the inspection system.</p><p><strong></strong></p><p><strong>Use case</strong></p><p>"As a manufacturer of automated inspection and sorting devices for agri-food, I want image analysis to assess the quality of every peanut kernel so I can remove damaged ones."</p><p></p><p><strong>Result</strong></p><p>Enabled per-kernel quality assessment and automated sorting by configurable damage thresholds.</p><p></p><p><strong>Expertise</strong></p><ul><li>Computer vision</li><li>Machine learning</li><li>Defect detection and classification</li><li>Dimensional measurement</li><li>Real-time sorting</li></ul>Fragen Sie Protech nach einem Angebot und erhalten Sie in kürzester Zeit eine Antwort.
Team
18 Mitglieder in Protech's Team
Senden Sie ihnen eine Nachricht und erhalten Sie in kürzester Zeit eine Antwort.
Bewertungen
Noch keine Bewertung für Protech
Teilen Sie Ihre Erfahrungen mit uns.
Kontakt
Kontaktdetails von Protech
Details
- HeadquarterVeselina Masleše 84, Novi Sad, Serbia
Building great business stories.
Werden Sie mühelos entdeckt, gewinnen Sie Projekte und wachsen Sie in über 500 Dienstleistungen. Sagen Sie uns, was Sie am besten können, und wir verbinden Sie mit den richtigen Kunden, um das Wachstum Ihrer Agentur zu fördern.
- Ein neues Projekt einstellen
- Erkunden
- Wie es funktioniert
- Preise
- Registrieren
- Blog
- Data Max
- Hilfe & Unterstützung
- Über
- Kontakt
- Jobs
- 2026 © Sortlist -
- Alle Rechte vorbehalten -
- Nutzungsbedingungen -
- Datenschutzbestimmungen
- - Datenverarbeitungsvertrag
- - Impressum
- - Verwalten Sie Ihre Cookies