t
talhaansari7

Muhammad Talha

@talhaansari7

Customer Satisfaction is my First Priority

Pakistan
Inglese, Urdu
Alcune informazioni sono riportate in lingua inglese.
Chi sono
I’m an AI Engineer and currently pursuing an MS in Artificial Intelligence and Autonomous Systems. I specialize in developing intelligent systems using machine learning, computer vision, and AI agent-based models. I’ve worked on real-world projects involving object detection, image classification, facial recognition, and automation. My skill set includes Python, OpenCV, PyTorch, TensorFlow, NumPy, Scikit-learn, Keras, Power BI, and Tableau. I’m committed to delivering scalable and innovative AI solutions tailored to your needs. Thank You!... Continua a leggere

Competenze

t
talhaansari7
Muhammad Talha
offline • 
Tempo di risposta medio: 1 ora

Consulta i miei servizi

Data engineer
I will do python data analysis and machine learning tasks on notebook and colab
A partire da45 USD / progettoMaggiori dettagli
Integrazioni IA
I will build genai, agentic ai, llm, rag apps, and ml,dl,CV projects
A partire da120 USD / progettoMaggiori dettagli

Esperienza lavorativa

Machine Learning Engineer

NUST • Full time

Jan 2025 - Present • 1 yr 4 mos

I am responsible for designing, developing, and deploying intelligent systems that can learn from data and make predictions or decisions without being explicitly programmed. Also focus on improving model accuracy, handling large datasets, and ensuring models can scale efficiently in production environments. Use programming languages like Python and tools such as TensorFlow, PyTorch, and scikit-learn. Further, collaborate with data analysts, software engineers, and business teams to translate data insights into practical applications.

Computer Vision Engineer

SubSeaScanning • Part time

Jun 2024 - Dec 2025 • 1 yr 6 mos

Conducted comparative analysis of point cloud registration (KISS-ICP, 3DTK, Open3D); 3DTK offered best efficiency for large-scale mapping, while KISS-ICP was 2× faster (handling 6.5° rotation & 10% overlap), validated with CloudCompare. Developing an underwater marker detection system by fine-tuning Detectron2; achieved 90% detection accuracy, extracted marker centers across RGB & grayscale channels (except red), supporting bundle adjustment optimization.