About
Machine Learning Engineer / Data Scientist Consultant with 8 years of experience in Data Science, Machine Learning, Generative AI, Agentic AI, and Advanced Analytics. Experienced in designing and deploying end-to-end ML and AI solutions, including LLMs, RAG applications, MCP-based integrations, NLP, and Time Series Forecasting. Proficient in Python, cloud platforms, and scalable AI architectures, with expertise in transforming structured and unstructured data into actionable business insights and enterprise-grade AI solutions.
Skills & Expertise (32)
Work Experience
Decision Science Consultant
Accenture
Jun 2025 - Present
Led the development of a Scenario-Aware GenAI Forecasting solution by combining traditional forecasting models with Generative AI-driven external signal analysis to improve forecast accuracy and business interpretability. Built and evaluated multiple forecasting approaches including XGBoost, LightGBM, LSTM, ANN, and ARIMA, selecting optimal models based on performance across different forecasting horizons and business scenarios. Integrated external market intelligence from news articles and industry reports using Large Language Models (LLMs) to capture disruption signals and emerging trends not present in historical data. Designed a Retrieval-Augmented Generation (RAG) pipeline leveraging embeddings and vector search to retrieve the most relevant contextual information and generate high-quality scenario summaries for forecasting. Developed automated feature extraction workflows using LLMs to convert unstructured documents and monthly reports into structured predictive variables for downstream forecasting models. Leveraged GCP Vertex AI, Model Garden, and enterprise GenAI capabilities to build scalable forecasting and insight-generation pipelines. Improved forecast performance by over 8% compared to the baseline forecasting framework while enhancing explainability through AI-generated scenario insights and business narratives.
Data Scientist / Machine Learning Engineer
Tredence Analytics INC
Mar 2023 - Jun 2025
Developed a Competitor Identification Framework, analyzing SKU-channel combinations to pinpoint top competitors. Identified the top 5 most capable competitors to our SKU-channel combinations across all 2042 SKUs vs. 5 channels. Monitored metrics to understand competitor behavior and strategize decisions based on their pricing impact on sales performance. Extracted valuable insights to inform pricing strategies, maintaining competitiveness and optimizing revenue effectively. Defined baseline sales as the starting sales level for a product or service, used as a performance reference point. Established baseline sales metrics to measure promotional impact and accurately gauge product performance. Differentiated baseline sales from promotional boosts to clarify revenue unaffected by promotions. Guided strategic decisions using baseline data to set goals, allocate resources, and assess promotional effectiveness. Enhanced performance evaluation by comparing actual sales with baseline figures, optimizing promotional strategies for increased revenue. Developed a dynamic pricing and promotion simulator to optimize sales strategies and profitability. Created dynamic dataframes to generate and manage future date projections for robust pricing analysis. Implemented baseline models for generating price recommendations and laying the foundation for advanced model comparisons. Enhanced decision-making processes by integrating automated insights into pricing strategies. Developed a dynamic pricing simulator that boosted sales by 15 percent through optimized pricing strategies.
Analyst / Data Scientist
InfoDeal Technologies
May 2018 - Feb 2023
Extracted restaurant data from AWS Redshift database through Kedro orchestration tool. Built a multi-label, multi-class classification model based on Deep Neural Network to predict cuisine code associated with restaurant ID and achieved the best accuracy score of 80.98 percent and Hamming loss of 0.6 percent on the validation data. Cleaned and tokenized the input data, vectorized the names, street, city, state, and country code, generating a document-word sparse matrix using TF-IDF, Tokenizer, and Count Vectorizer. Experimented with parameters such as min-df, max-df, token-pattern, and n-gram range for optimization. Experimented with multiple classification algorithms such as Label Powerset, Multi-Learn KNN, LSTM, and OneVsRestClassifier. MLOps Activities: Fetched data from Redshift database in AWS Sagemaker Notebook instance, preprocessed and feature-engineered new restaurant name, address, city, and description into ”Detailed SIC”. Saved versioned training tables into Feature Store, built LSTM model, and registered it in the Model Registry. Created endpoint and deployed LSTM version to the endpoint for batch predictions.
Education
Bachelor of Engineering - Mechanical Engineering - Pune University
- · Afghanistan
Certifications
No certifications added yet
Interested in this developer?
Profile Score Breakdown
Profile Overview
Skills (32)
Click a skill to find developers with the same skill
