Sampreeth Venkatasubramanian

I am a highly motivated individual with a strong desire to learn new technologies and continuously expand my skill set. With well-developed problem analysis, problem-solving, and issue resolution skills, I am adept at identifying challenges and delivering effective solutions. My ability to communicate complex technical concepts clearly and professionally to both technical and non-technical audiences ensures seamless collaboration and understanding.

I possess strong organizational skills and meticulous attention to detail, allowing me to manage multiple priorities and deadlines effectively. I am equally comfortable working independently or as part of an agile development team, bringing my Army National Guard experience as a CBRN specialist to emphasize discipline, teamwork, and strategic thinking. I pride myself on my ability to adapt, prioritize, and communicate challenges in a timely manner, ensuring successful project outcomes.

• Selfie continuous sign language recognition using neural network, D. A. Kumar, P. V. V. Kishore, A. S. C. S. Sastry, P. R. G. Swamy, 2016 IEEE Annual India Conference (INDICON), 2016, 1-6, 10.1109/INDICON.2016.7839069
• Explainable Deep Learning, A library for generating explanations for Generative AI for textual and multimodal inputs. Performed fine-tuning for LLMs and Vision Language Models. Visualized Data Influence using Captum.
• SalASR, Saliency-highlighting in Automatic Speech Recognition (ASR) for fine-tuned Whisper models.
• RAG, Developed a RAG system for Document-based QA bot for customer support using Langchain with ChromaDB, LLama-2 and Mistral models. Implemented prompt optimization for agentic environments using DSPy.

Artificial Intelligence, Deep Learning, Generative AI, Machine Learning, Reinforcement Learning, Data Science, Product Management, MLOps, Agile, Data Analysis, DevOps, Data Visualization, Python, Bash, SQL, Java, PyTorch, TensorFlow, C/C++, Pandas, Numpy, Scikit-Learn, LangChain, OpenAI, DSPy, LLaMA, Mistral, Joblib, DeepSpeed, FAISS, Flask, Git, Docker, Linux, AWS, GCP

Relocation: Open to relocation within the US

• Received certification through the US Army in November of 2021.