Anirban Deb

Investigated foundational and architectural principles of Liquid Neural Networks, emphasizing temporal adaptability for dynamic environments.

Modified Liquid Time-constant Networks (LTCs) to improve performance, stability, and efficiency in dynamic environments, enhancing system robustness.

Implemented adaptive learning with feedback mechanisms to enhance behavior in sequential, dynamic, noisy scenarios.

Contributed to research introducing novel algorithmic advancements in liquid networks for real-time applications, pushing the boundaries of AI research.

Built Transformer-based (BERT) and TF-IDF models to extract semantic meaning from diverse unstructured textual data, enhancing data analysis capabilities.

Fine-tuned Large Language Models (LLMs) for domain-specific Natural Language Processing (NLP) inference and classification, improving model accuracy.

Designed robust and secure REST APIs using FastAPI, validating model endpoints and optimizing for latency, scalability, and deployment efficiency.

Developed Sentiment Analysis, Attention Mechanisms, and semantic embeddings utilizing real-world datasets with annotated labels for enhanced model performance.

Conducted detailed simulations and system optimization using HOMER Pro to evaluate cost-effective hybrid configurations, identifying optimal energy solutions.

Explored advanced integration of Reinforcement Learning algorithms for optimizing component selection and operational cost in HRES, enhancing system efficiency.

Analyzed diverse real-world energy data to assess feasibility, storage needs, and environmental performance of HRES models, informing strategic decisions.