Jan 2024 – Feb 2025 · 1 yr 1 mo

• As a Data Scientist at Xoriant, I focus on leveraging Generative AI and Large Language Models (LLMs) to develop state-of-the-art chatbot solutions. My key responsibilities and accomplishments include:
• Advanced Chatbot Development:
• Created a highly efficient chatbot using the "Chain of Thoughts" methodology, enhancing logical reasoning and contextual understanding.
• Integrated Retrieval-Augmented Generation (RAG) with internet search capabilities to improve the chatbot's response accuracy and relevance.
• Experimented with various LLMs, including implementing the llama Index React Agent and developing custom agents tailored to specific use cases.
• Built another sophisticated chatbot using Azure services, incorporating RAG with Azure Cognitive Search (ACS) and Azure Prompt Flow to streamline information retrieval and response generation.

Aug 2023 – Jan 2024 · 5 mos · Bengaluru, Karnataka, India

• Designing, developing, and optimizing stored procedures to enhance backend functionality of applications.
• Debugging and troubleshooting database-related issues to maintain data integrity.
• Performance tuning of stored procedures and SQL queries for optimal database performance.

Dec 2021 – Jun 2023 · 1 yr 6 mos · Pune, Maharashtra, India

• Data Augmentation using Deep Learning:
• Generated synthetic data for modeling using ML/DL techniques.
• Employed Deep Learning Algorithms to create synthetic datasets.
• POC - Data Discovery through NLP:
• Extracted sentences with similar semantic meanings from datasets.
• Utilized Spacy library for word embedding to meet business objectives.
• Image Captioning using Deep Learning with Attention Model (An Eye for Blind People):
• Developed an image captioning model and integrated text-to-speech for accessibility.
• Preprocessed captions using NLP and Keras.
• Built an Encoder-Attention-Decoder model with CNN for feature extraction and RNN for generating captions.
• Achieved BLEU score for evaluating caption predictions.
• Employed Google text-to-speech API for reading captions aloud.

Nov 2020 – Dec 2021 · 1 yr 1 mo · Pune, Maharashtra, India

• Responsible for creating POCs (Proof Of Concepts) using AI/ML algorithms to enhance iEDPS (Data Privacy Solution) product.
• Conducted exploratory data analysis on 40k Amazon reviews for sentiment analysis, visualization, and data manipulation.
• Applied NLP techniques such as stop words removal, lemmatization, and tf-idf vectorization for corpus preprocessing.
• Developed models using ML algorithms including MultiNB, MultiSVM, Neural Networks, and LSTM.
• Implemented over-sampling with imbalanced-learn library for dataset balancing.
• Utilized ensemble techniques like XGBoost with MultiNB to achieve maximum F1 score.
• Created confusion matrix for model performance evaluation.
• Performed topic modeling after clustering using LDA.