Accurate, meticulous, and result-oriented AI Specialist with a proven track record in deploying advanced Generative AI, Deep Learning, and Machine Learning techniques to identify patterns, extract valuable insights, and solve complex business challenges. Experienced in planning and executing diverse AI-driven projects, collaborating with key stakeholders to define problem statements and deliver impactful, data-driven solutions.
Overview
11
11
years of professional experience
1
1
Certification
Work History
ML Senior Software Engineer
Iron Mountain
08.2023 - Current
Designed and implemented a state-of-art Generative AI solution for Document Classification, Extraction and Question answering, leveraging both closed ( GPT-4, Gemini, and Claude) and Open( Mistral) to achieve high accuracy and efficiency
Handled both taxonomy and non-taxonomy use cases for classification and extraction across different document types.
Implemented a Retrieval-Augmented Generation (RAG)-based Question-Answering solution using LLMs, seamlessly integrated with MongoDB Atlas for semantic search. The system utilized OpenAI embeddings to encode documents into vector representations, which were stored in MongoDB Atlas Search for efficient similarity-based retrieval.
implemented an intelligent document splitting solution using LLMs, which automatically identifies logical sections within documents for accurate classification and segmentation. This feature supports large files (>1000 pages) by splitting them into manageable chunks, ensuring compatibility with LLM’s context window without losing critical context between splits. The solution dynamically handles a wide range of document types and structures, ensuring precise and scalable document processing.
Designed and implemented a document reduction service using MS Presidio, integrated with spaCy and Robusta models for entity recognition. Supported both text and PDFs, allowing users to define entities with allow_list, deny_list, and regex-based patterns. Integrated Google OCR for improved text extraction and added custom EntityRecognizers for specific use cases. Provided various redaction options, including complete masking, ensuring compliance with data privacy regulations.
Data Science Delivery Manager
Acuity Knowledge Partners
09.2022 - 06.2023
Developed and implemented advanced time series forecasting models using machine learning algorithms, such as ARIMA, Prophet, and LSTM, to accurately predict future sales trends for a Japanese luxury omnibrand curator and distributor
Collaborated with cross-functional teams, including sales, marketing, and supply chain, to gather insights and inputs on market trends, customer behavior, and product availability that could impact future sales
Utilized statistical metrics MAE to measure and evaluate the accuracy of the forecasting models, identifying areas for improvement and implementing necessary adjustments to increase model performance
Contributed to the development of a comprehensive demand planning process that integrated sales forecasting, inventory optimization, and supply chain management to minimize stockouts and reduce inventory holding costs by $1.4B per year
Led a data science team in developing a proprietary product that analyzed financial statements, earnings call transcripts, holdings data, and news articles to uncover trends and potential red flags affecting stock performance. Utilized transfer learning with pre-trained models like BERT for sentiment analysis of stakeholder statements and implemented topic modeling on earnings call transcripts to identify key themes and cluster documents by products and industries. Delivered actionable insights via Power BI dashboards, empowering investors and equity analysts with data-driven strategies.
Senior Data Scientist
Harman Connected Services
06.2019 - 06.2022
Developed a sophisticated revenue forecasting model using PySpark and Python for a US market research company to identify the price elasticity of demand for different products at a Store-Week level
Conducted thorough data cleaning and preprocessing to ensure the accuracy and consistency of the data used in the forecasting process
Employed log-log regression model, which is known for its robustness and accuracy in handling non-linear relationships, to analyze the impact of various external variables, including price, discount, seasonality, marketing spend, and competitor distribution, on revenue growth
Utilized cross-validation techniques, such as k-fold validation, to test the model's accuracy and generalization performance
Developed market mix models using Python to analyze the impact of various marketing levers, including price, promotion, advertising, and distribution, on sales performance for a retail company
Preprocessed and transformed the data to account for seasonality and trends, and to incorporate the adstock and adbudget parameters into the modeling process
Utilized log-log regression models to capture the non-linear relationships between the marketing levers and sales, and estimate the price elasticity of demand for different products at the store-week level
Incorporated advanced feature engineering techniques, such as interaction terms, lags, and dynamic effects, to capture the complex dynamics of the market and improve the model's predictive power
Implemented a switch matrix in conjunction with chi-square test to check if there is a significant difference in the distribution of brand switching between two groups
Conducted data analysis to identify patterns in brand switching behavior and identify factors that influence brand loyalty
Provided insights into consumer preferences and purchasing behavior to optimize marketing and advertising strategies
Developed a model using NLP to identify and analyse the trend in consumer sentiments towards a brand and compare with category
Provided insights into the impact of social media and other digital channels on brand perception and reputation
Designed and created NLP based model using custom NER (Named entity recognition) and CRF (Conditional Random Fields) to identify diseases and the corresponding treatments to recognize entities in healthcare data for a US based pharmaceutical company
HCS is a pioneer in providing AI&ML-enabled solutions across retail, banking, insurance and telecom industries
DATA ANALYST
IBM India
12.2013 - 06.2019
Conducted exploratory data analysis (EDA) on large-scale telecom customer data to understand customer behavior and identify key factors that contribute to churn
Developed and implemented machine learning models, including logistic regression, decision trees, and random forests, to predict customer churn probability and identify at-risk customers
Conducted feature engineering techniques such as PCA and feature scaling to improve model performance and accuracy
Utilized classification metrics such as precision, recall, F1-score, and AUC-ROC to evaluate the performance of the models and identify areas for improvement
Component owner for order management
Conducting data mining, data modeling and report generation for the UK based telecom companies Inventory management based on the Customer orders and the Real Time Availability Monitoring (RTAM) reports
Consulting with clients and team members to document and understand business and technical requirements for Order management system functions provided by software platform
Train and mentor new joiners on Quality inspection, product specification and company policies
IBM offers business and technical consulting, training and support, data storage, analysis, networks, and enterprise applications across different clients worldwide
Education
Master of Technology - Microelectronics
NIT Silchar
05.2013
Bachelor of Technology - Electronics and Communication