Google Certified AI/ML Engineer with 8+ years of experience in Machine Learning, Deep Learning, Generative AI, Time Series Analysis, Statistics, Big Data, DevOps, MLOps, and advanced analytical techniques. Skilled in delivering end-to-end AI and data science solutions, encompassing objective discovery, data preprocessing, model development, deployment, monitoring, and continuous optimization.
Expertise in the full life cycle of AI/ML algorithms: Extensive experience in all stages, including objective discovery, data preprocessing, training and iteration, validation, performance improvement, deployment, monitoring, and algorithm updates.
Machine Learning & Deep Learning: Mastery of algorithms such as Linear Regression, Logistic Regression, Random Forest, XGBoost, LightGBM, CatBoost, Naïve Bayes, KNN, SVM, PCA, LDA, K-Means, K-Modes, Time Series Analysis, NN, RNN, and LSTM.
NLP & Generative AI: Text Classification, Word Embedding, NER, Topic Modelling, RAG, Transformer (BERT, Llama2, Gemini) and Vector Databases (Chroma DB).
ML-Ops: Proficiency with tools and best practices, including Git, Kubeflow/ML Flow, Docker, Kubernetes, CI/CD pipelines, and Airflow.
Problem-Solving: Proven analytical, mathematical, and creative problem-solving skills with effective task prioritization and execution.
Communication: Excellent professional communication skills with stakeholders and business colleagues.
MTM (Meter Transformer Mapping): Implemented advanced data conditioning techniques, time series k-means clustering, and a statistical algorithm to detect incorrect meter-to-transformer mappings and generate recommendations for accurate transformer assignments
PCEM (Predictive Cell Energy Management): Optimized radio technology usage by dynamically locking and unlocking it during low-traffic periods for 4G/5G cells, leveraging AI/ML predictions. This approach ensures minimal impact on customer experience while saving energy on radio equipment, reducing telecom vendors' operational costs.
IBO (Intent-Based Operation): Analyzed user intent and performed real-time utilization, AI/ML forecasting for 5G partitions/slices. The system autonomously adjusts network utilization to meet intended requirements if the forecasted data does not align with user intent.
1. E&I to M&R Migration Journey: Analyzed historical data to capture the journey of members migrating from Product E&I to M&R.
2. Built a response model to predict members likely to migrate from E&I to M&R, boosting marketing campaign efficiency by increasing responses, or reducing expenses.
3. Customer Feedback Analysis: Developed a predictive model to detect the most sarcastic comments, as well as partially to highly negative comments.
NER Model on Airlines Consumer Comments: Built an NER model to identify source and destination cities from airline consumer comments for one of the largest airline services.
Data Analysis on Health Data of Mining Machines: Part of a team that developed a platform connecting mining equipment using IoT devices to enable real-time monitoring and effective utilization of mining equipment