Dhanashri Shivdas

Recent B.Tech graduate from IIT Delhi (Electrical Engineering, CGPA 8.12) specializing in ML systems, LLM applications, and Agentic AI. Built and deployed end-to-end AI projects including an agentic hiring assistant, hybrid emotion detection system, and EEG-based robot control pipeline with proven production metrics. github

Machine Learning Research Intern at Uppsala University — built an audio fingerprinting pipeline using FFT spectrograms and k-NN classification achieving 90% accuracy. Also interned at Trident Group as Software and Data Engineering Intern, building Python pipelines for 10,000+ records achieving 92% accuracy and reducing fault response latency by 33%. github+1

Key projects include YOLOv8-based drone detection with GAN-augmented dataset achieving 93% accuracy, EEG-based Brain Computer Interface for real-time robot control achieving 89% accuracy, and an LLM agentic hiring system using LangChain and Groq APIs reducing screening effort by 40%. github+1

Currently seeking a 2-3 month research internship (June-August 2026) in AI/ML — open to remote globally and onsite in Germany, South Korea, or Japan — before starting full-time in September 2026; actively learning AI security and adversarial ML. github

Technical skills: Python, PyTorch, LangChain, LLMs, Agentic AI, RAG pipelines, Generative AI, Computer Vision, Signal Processing, Docker, Streamlit, FastAPI, CUDA, HuggingFace. github