Zum Hauptinhalt springen
Preet Koshiya

Preet Koshiya

Full stack developer, Software engineer

Talente
Deggendorf, Landkreis Deggendorf, NiederbayernMitglied seit 13. August 2024
·

Diese Person einstellen

Senden Sie diesem Kandidaten direkt ein Jobangebot

Über

Ich studiere KI an der Technischen Hochschule Deggendorf unter der Leitung von Prof. Patrick Glauner. Ich habe verschiedene neue Fächer gelernt, wie Algorithmen und Datenstruktur, KI-Projekte im Rahmen eines B-Plus-Unternehmens, Computer Vision, Computerlogik, Statistik und Projektmanagement. Ich liebe das Programmieren und komme gut mit der Programmiersprache Python zurecht.

Berufserfahrung

ODD Recognition (2024) under Prof. Patrick Glauner

  • Engineered an Odd Recognition model improving precision by 15% through advanced pre-processing techniques.
  • Utilized F1 score metrics for comprehensive model evaluation
  • Achieving an accuracy of approximately 79.5% demonstrates the model’s effectiveness in distinguishing between different weather categories such as clear, rainy, snowy and foggy.

Ausbildung

University Project Deggendorf, Bavaria Android App Development (Mar 2023 - Jun 2023)

  • Developed and deployed a comprehensive cooking recipe app on Android.
  • Improved app performance by optimizing backend algorithms by 20%.

Face Detection (Mar 2023 - Jul 2023)

  • Created a Python-based face recognition system using OpenCV and TensorFlow.
  • Implemented real-time facial recognition with an accuracy rate of 95%.

Speech Recognition using Pydub Library (Oct 2022 - Feb 2023)

  • Built a speech-to-text converter with enhanced noise reduction capabilities.
  • Achieved a recognition accuracy of 90% in various acoustic environments.

Data Exploration and Prediction for House Prices (Nov 2023 – Dec 2023)

  • Conducted extensive data analysis and modeling using Python libraries.
  • Predicted house prices with an R² score of 0.85.

Data Analysis and Machine Learning Model of Water Analysis using Support Vector Classifier (Jan 2023)

  • Developed a classification model with an accuracy of 82% for water quality analysis.
  • Presented findings at the university's annual tech symposium.

Bewertungen