ADITHYA BOYAPATI

Utilize Deep Learning and Generative AI models to create cutting-edge solutions for intricate business challenges in the IT sector,specifically within Data Science, Engineering, and Machine Learning domains.

• Data Scientist with extensive experience in developing sophisticated data-driven solutions and addressing complex business challenges with minimal supervision.
• Deep expertise in Deep Learning and Generative AI, specializing in advanced architectures like CNNs, RNNs, and LLMs (e.g., GPTs, LLaMA), as well as diffusion models such as Stable Diffusion and DALL·E.
• Proficient in deploying and scaling Machine Learning models on cloud platforms like Azure ML, driving business innovation and operational efficiency.
• Adept at building custom contextual search and summarization systems using fine-tuned LLMs, streamlining information retrieval from organizational documents.
• Experienced in mentoring and leading AI teams, proposing robust solution architectures for diverse projects, and actively contributing to recruitment initiatives to foster team growth.
• Strong background in predictive analytics and advanced model optimization techniques, including quantization, pruning, and knowledge distillation, to improve model performance and efficiency.
• Recognized for consistent excellence in project delivery, earning accolades such as Performer of the Quarter at Evoke Technologies and a Certification of Appreciation from Accenture.
• Hands-on experience with a wide range of technologies and tools, including Python libraries (PyTorch, TensorFlow), Business Intelligence tools (Power BI, Tableau), and databases (Oracle DB, MySQL).
• Demonstrated ability to design and implement impactful solutions, such as a custom travel chatbot and customer attrition models, achieving measurable improvements in business processes and decision-making.

Collaborator  Problem-solving   Communicator   Critical Thinking  Attention to Detail

Generative AI     Deep Learning    Machine Learning    Predictive Analytics

Model Optimization    Data Analysis and Reporting    Cloud Platforms (Azure ML)

Large Language Models (LLMs)     Custom Search & Summarization 

Natural Language Processing(NLP)

• Performer of the Quarter [Evoke Technologies]: Acknowledged for exceptional performance and dedication, remarkably contributing to project deliveries.
• Certificate of Achievement in Data Analytics[Ineuron]: Completed a Data Analytics course with Distinction.

Project Title: Custom Travel Chatbot

Technologies Used: Azure AI Studio, Azure Bot Service, Microsoft Teams, Azure Cognitive Services, Embeddings

Key Result Areas:

• Developed an advanced "Custom Travel Chatbot" by leveraging a fine-tuned LLM (Language Model) for extracting contextual search results and summarizing information from internal company documents to design the overall architecture of the chatbot.
• Utilized Azure AI Studio to build and fine-tune Large Language Models specific to travel policies, enhancing the chatbot's natural language understanding and response generation.
• Created custom embeddings for travel documents using Azure Cognitive Services to facilitate accurate information retrieval and presentation.
• Integrated the chatbot with Microsoft Teams through Azure Bot Service, configuring settings such as endpoint URLs and messaging formats to ensure compatibility with Teams' API and messaging standards.

Project Title: Extracting Entity Information from Outlook Mails using OpenAI

Technologies Used: Python, Alteryx, OpenAI API, GPT Models (GPT-3 Text-Davinci-003), Text Preprocessing Libraries (NLTK, SpaCy)

Key Result Areas:

• Created a system to extract entity-related information from Outlook emails received from government websites across various US jurisdictions using the OpenAI API.
• Utilized Alteryx to access and retrieve email data from Outlook, ensuring secure and authorized access.
• Applied text preprocessing techniques, such as tokenization, stop-word removal, and normalization, to clean the extracted text and prepare it for entity extraction.
• Integrated OpenAI's API to perform natural language processing tasks, leveraging GPT models to extract entities from email text.
• Developed custom prompts to guide the OpenAI model in accurately recognizing and extracting relevant entities and fine-tuned the model with domain-specific data to enhance performance.

Project Title: Customer Attrition Prediction

Technologies Used: Alteryx, Python, Pandas, NumPy, scikit-learn, Seaborn, Matplotlib, GridSearchCV, ML Algorithms, Flask, REST API, CRM Systems, Data Pipelines

Key Result Areas:

• Developed and optimized a Customer Attrition model using predictive analytics techniques.
• Aggregated customer data from multiple sources and conducted comprehensive exploratory data analysis (EDA) using BI tools.
• Evaluated various Machine Learning algorithms, including Logistic Regression, Random Forest, and Gradient Boosting, to identify the best model for churn prediction.
• Trained the selected model using scikit-learn and cross-validation techniques to ensure robustness and prevent overfitting.
• Analyzed feature importance to identify key factors contributing to customer churn, providing actionable insights for targeted interventions.
• Deployed the trained model into production using Flask to create an API endpoint, and developed a comprehensive strategy based on model insights, including personalized retention campaigns, targeted offers, and enhanced customer support interventions.

Project Title: Customer Late Payment Prediction

Technologies Used: Alteryx, Python, Pandas, scikit-learn, Seaborn, Matplotlib, ML Algorithms, GridSearchCV, Flask, REST API, Financial Systems, Data Pipelines

Key Result Areas:

• Implemented a predictive model for customer late payments using advanced analytics and Machine Learning algorithms to analyze historical payment data and identify patterns for robust prediction models.
• Analyzed correlations between features such as payment history, customer credit scores, and late payment occurrences to determine significant predictors.
• Evaluated various Machine Learning algorithms, including Logistic Regression, Decision Trees, and Random Forest, to select the best model for late payment prediction.
• Deployed the final model using Flask to create an API endpoint, allowing seamless integration with the company's financial system for real-time late payment predictions.
• Implemented monitoring tools to track model performance and accuracy over time, ensuring reliable predictions.

Project Title: Customer Product Recommendation System using Unsupervised Learning

Technologies Used: Alteryx, Python, Pandas, NumPy, scikit-learn, Clustering Algorithms (K-means Clustering), Power BI, CRM Systems, Oracle DB

Key Result Areas:

• Designed and implemented a recommendation system to suggest products to customers based on their historical purchase patterns and business behaviors, utilizing unsupervised learning algorithms such as clustering.
• Collected customer transaction data, including purchase history, product interactions, and business behavior metrics from sources such as CRM systems and e-commerce platforms.
• Identified key features such as purchase frequency, recency, product categories, and customer demographics to capture customer behavior and preferences.
• Implemented clustering algorithms, such as K-means Clustering, to segment customers based on similarities in purchase patterns and behaviors.
• Analyzed and profiled each cluster to understand the characteristics and preferences of customers within each segment and tailored recommendations based on these insights.

Project Title: Revenue Forecasting

Technologies Used: Time Series Models (ARIMA, SARIMA, LSTM), Python, Pandas, Matplotlib

Key Result Areas:

• Built a robust forecasting model to predict revenue across sales regions and billing entities, covering over 30 renewal services and 8 transactional services.
• Conducted extensive model evaluation using statistical and deep learning methods, with the LSTM model achieving superior accuracy and the lowest Mean Absolute Percentage Error (MAPE).
• Automated data preprocessing pipelines, including seasonality and trend decomposition, to streamline model training.
• Delivered dynamic dashboards to visualize revenue trends and forecast results for improved business decision-making.
• Gained accolades from senior leadership for delivering highly accurate forecasts, enabling better resource allocation and revenue planning.
• Implemented a feedback loop for continuous model improvement based on updated sales and revenue data.