Service Advisor

Call BA - Subsidiary of British Airways

June 2024 - Present

• Responsible for investigating fraudulent transactions and facilitating chargebacks in collaboration with banks.
• Handling ticket issuance and reissuance in accordance with fare regulations using Amadeus.

Senior Tech Specialist

Ienergizer IT Services Private Limited

November 2022 - May 2024

• Played a key role in elevating the user experience for a diverse global player base for a leading US gaming company.
• Achieved measurable improvement in customer satisfaction through efficient issue resolution and strategic troubleshooting.
• Served as a game moderator, ensuring adherence to community guidelines and fostering a safe, inclusive, and respectful gaming environment.
• Safeguarded user data and upheld account security in compliance with international data privacy regulations (GDPR, CISPA, COPPA, LGPD).
• Actively contributed to the prevention of financial fraud and cyber threats by enforcing policies aligned with these standards.

IBM Data Science Professional Certificate

Coursera

Provides an excellent broad beginner foundation covering the entire data science lifecycle, tools (Python, SQL, Jupyter, Pandas, Scikit-learn basics, IBM Watson Studio), and methodologies.

Applied Data Science with Python Specialization

University of Michigan, Coursera

Deepens Python skills for data science, covering advanced library usage (Pandas, Matplotlib, Scikit-learn in-depth, NLTK) and problem-solving techniques.

Deep Learning Specialization

DeepLearning.AI (Andrew Ng), Coursera

Considered a gold standard for learning deep learning fundamentals (CNNs, RNNs), covering theory and practical implementation in Python, TensorFlow, and Keras.

Hands-on Foundations for Data Science and Machine Learning with Google Cloud Labs Specialization

Coursera

Directly bridges foundational DS/ML knowledge to the GCP ecosystem with practical labs using key GCP data and ML tools like BigQuery, Vertex AI Notebooks, Cloud Storage, and basic Vertex AI (AutoML, Training).

Data Engineering, Big Data, and ML on GCP Specialization

Google Cloud, Coursera

Core for understanding how to design, build, and manage scalable data pipelines and data processing systems on GCP (BigQuery, Dataflow, Pub/Sub, Dataproc, Cloud Composer).

Machine Learning on Google Cloud Specialization

Google Cloud, Coursera

Focuses on cloud-native ML practices, leveraging the full suite of Vertex AI (Training, Prediction, Endpoints, Pipelines, Feature Store, AutoML) and BigQuery ML with TensorFlow/Keras.

Advanced Machine Learning on Google Cloud Specialization

Coursera

Covers advanced ML topics like NLP with Transformers (BERT), Computer Vision, Recommenders, and MLOps practices (Vertex AI Pipelines, TFX, Explainable AI) for productionizing models on GCP.

IBM Generative AI Engineering Professional Certificate

Coursera

Provides a structured approach to learning about Large Language Models (LLMs), prompt engineering, and building applications with Generative AI (Python, Hugging Face, LangChain, Vector DBs, LLM APIs like Vertex AI Gemini).

Google Cloud Professional Machine Learning Engineer

Preparing via Coursera & Google Cloud Skills Boost

Aims to validate expertise in designing, building, and productionizing ML models on Google Cloud, covering all aspects of the ML lifecycle.

Google (via Coursera)

Developed foundational skills in data collection, transformation, analysis, visualization (e.g., Tableau, R), and data-driven decision-making using spreadsheets and SQL.

Google (via Coursera)

Built upon foundational data analytics with advanced topics such as statistical analysis (e.g., regression), an introduction to machine learning, and advanced data visualization techniques using Python.

Microsoft

Gained proficiency in using Power BI to transform, model, visualize, and analyze data, enabling the creation of impactful reports and dashboards for data-driven insights.

Imperial College London (via Coursera)

This certification helped me to learn mathematics for machine learning, focusing on linear algebra, multivariate calculus, and principal component analysis (PCA). Also, I learned about vector spaces, matrices, eigenvalues, optimization techniques, and dimensionality reduction. It strengthened my ability to apply mathematical reasoning to ML algorithms and data representation..

LLM Concepts RAG System

Elasticsearch, Streamlit, Docker, Groq, OpenAI | GitHub Link

• Advanced Retrieval-Augmented Generation for Machine Learning Education
• An end-to-end RAG application that answers ML questions using hybrid search (BM25 + Vector) and cross-encoder reranking.
• It features dual-LLM fallback (Groq & OpenAI), dynamic Mermaid diagram generation for visual learning, and a fully Dockerized architecture with PostgreSQL & Grafana for monitoring interactions.

SaarthiEO — Disaster Detection from Satellite & Drone Imagery

Python, PyTorch, Gradio, Hugging Face | GitHub Link

• Fine-tuned EfficientNet-B3 on a 4-class aerial disaster dataset achieving 97.38% test accuracy (Fire, Flood, Collapsed Building, Traffic Incident).
• Built an end-to-end Grad-CAM explainability pipeline hooking the last convolutional block to produce interpretable attention heatmaps.
• Deployed a production-grade 3-tab Gradio web app (single predict, Grad-CAM, batch CSV export) to Hugging Face spaces with Git LFS model storage.

aetherRead — AI-powered Offline-First PDF Research Workspace

Python, Next.js, TypeScript, Ollama, ChromaDB | GitHub Link

• Designed and engineered a cross-platform, production-grade PDF reader prioritizing privacy and offline-first performance, ensuring zero data dependency on external cloud services.
• Implemented a local-first storage architecture utilizing IndexedDB (Dexie.js) for high-performance management of PDFs, reading state, and user-generated annotations.
• Architecting a local RAG pipeline using Ollama (llama3.2:3b) and ChromaDB to enable intelligent document querying, semantic search, and AI-assisted research with verifiable citations.
• Engineered a "Comfort Engine" featuring six customizable reading themes and page-level annotation capabilities categorized for research workflows (Important, Definition, Question, Quote).

Customer Churn Prediction & Retention Prioritization

Python, XGBoost, SHAP, Power BI | GitHub Link

• Focus: Churn prediction, A/B testing, behavioral segmentation for proactive customer retention
• Built churn prediction model (XGBoost, 87% AUC) using transactional and behavioral features from 10K+ credit card customers
• Applied SHAP values to identify top 5 churn drivers (e.g., declining transaction frequency, product disengagement); created risk-based prioritization tiers (High/Medium/Low) for targeted interventions
• Developed Power BI dashboard enabling sales teams to filter at-risk accounts by value tier, supporting proactive retention workflows
• Designed reproducible pipeline with synthetic data generator and automated evaluation; documented limitations and monitoring considerations for production readiness