Karan Anchan

Master's Student in CS (Artificial Intelligence)
📍 Freiburg im Breisgau, Germany

👋 About Me

I am a specialized AI Researcher & Engineer currently pursuing my Master's at the University of Freiburg. My focus lies at the intersection of Generative AI, Computer Vision, and NLP. I engineer end-to-end pipelines that solve real-world problems—from medical image segmentation to semantic translation systems.

🚀 Featured Projects

🏥 UNETR: 3D Abdomen Segmentation

Stack: PyTorch MONAI Vision Transformers

I architected a volumetric segmentation pipeline using the UNETR (Vision Transformer) architecture to identify 13 distinct abdominal organs. This project solves the challenge of capturing global context dependencies in medical imaging.

Fig 1: The UNETR Architecture implementing Transformer Encoders for volumetric data.

Key Achievements:

🏆 0.8027 Mean Dice Score achieved on the BTCV dataset.
Preprocessing: Engineered robust HU windowing and isometric resampling to normalize soft-tissue density.

Fig 2: Actual model output showing precise segmentation of internal organs.

View Project on GitHub

🗣️ Neural Machine Translation (Eng-Hi)

Stack: PyTorch Transformers Altair

A complete implementation of the Transformer architecture ("Attention Is All You Need") built from scratch. This model handles complex transliterations and formal grammar with high accuracy.

Performance Metrics:

📈 SacreBLEU: 62.48
📉 Character Error Rate (CER): 0.28 (28%)
🤖 CHRF++: 71.78

Fig 3: Training progression showing correlation between SacreBLEU scores and error reduction.

View Project on GitHub

💼 Professional Experience

Machine Learning Intern | Wizdom Ed

Mangalore, India | Oct 2023 – Apr 2024

Engineered a Retrieval-Augmented Generation (RAG) pipeline using LangChain to relate LLM responses to curriculum data.

🚀 35% increase in student engagement via optimized data ingestion.
✅ 30% boost in positive response rates by implementing retrieval quality feedback loops.

🎓 Education

M.Sc. Computer Science (AI)
Albert Ludwig University of Freiburg

Apr 2025 – Present
Grade: 2.5 (Gut)

B.E. Computer Science
N.M.A.M Institute of Technology

Sep 2020 – Aug 2024
CGPA: 9.33/10