Karthik Pagnis

Machine Learning Engineer | Generative AI · RAG · LangGraph | Computer Vision | PyTorch · Azure ML | IIT Madras M.Tech AI | Open to Full-Time Roles

SanDisk (via Magnit) Bengaluru, India

Design Verification Engineer | ML, Generative AI & Computer Vision Jul 2024 – Mar 2026

• Architected an NLP/ML text classification system for error log processing, reports, and test specs via SpaCy, NLTK, and fine-tuned models (Hugging Face, PyTorch); cut manual triage by 60% and earned the Trade Secret Award (2025).
• Engineered a RAG-powered debug assistant using LLMs (Llama, Mistral) and LangChain, enabling engineers to query past failure reports, test cases, and internal wikis in natural language to surface root causes; reduced product launch timelines by 15%;

deployed via Azure ML and MLflow with CI/CD.

• Built an end-to-end wafer defect detection system for NAND flash inspection using PatchCore (unsupervised anomaly detection)

and fine-tuned EfficientNetB0 (supervised classification), with Grad-CAM for defect localization; deployed as a FastAPI + Docker service on Hugging Face Spaces, achieving 94%+ AUROC on the MVTec industrial anomaly benchmark and reducing manual inspection effort by∼60%.

Western Digital Bengaluru, India

Machine Learning Engineer Intern Oct 2023 – Jun 2024

• Developed pattern recognition and anomaly detection models (Scikit-learn, XGBoost) to identify recurring failure signatures and group similar failures across thousands of test runs; automated 40% of manual validation tasks and saved 15+ engineering hours weekly.

Indian Institute of Technology Madras (IIT Madras) Sep 2025 – Jun 2027 (Expected)

M.Tech in Artificial Intelligence Online

REVA University, Bengaluru Dec 2020 – Jun 2024

B.Tech in Artificial Intelligence & Machine Learning