.png&w=256&q=75)
0
0
India
2+ years of experience
We have extensive experience building AI systems across multiple domains, including multi-agent systems, LLM tooling, computer vision, and applied research: Interned at Intel as Project Intern, where we pipelined and benchmarked Intelβs DLStreamer framework for real-time object detection, focusing on optimization and performance. Collaborating with IISER on research related to hallucination quantification in LLMs and improving explainability across legal, medical, and financial domains. Built a ReAct-based multi-agent system using Autogen with a knowledge graph memory (Zep) and Microsoft Presidio for data anonymization. Developed AI tools at Samsung R&D for video object segmentation and annotation using XMem++, with an interactive memory-based feedback loop. Created a federated recommendation system using PySyft and Flower for privacy-preserving collaborative learning across distributed clients. Applied RAG (Retrieval-Augmented Generation) techniques in legal tech and bug triaging tools to enhance semantic search and document grounding. Built deep learning models from scratch (SVMs, GANs, CNNs) and optimized object detection pipelines for real-time inference. Worked on applied research projects in phishing detection, prostate cancer forecasting, and wildlife monitoring using domain-adapted models. Our team is well-versed in tools like LangChain, PyTorch, Hugging Face, and GCP, and trained in ML, DL, and NLP through multiple certifications.