Aswath Ramana

Over 6 years of experience in machine learning, AI, and data science, specializing in end-to-end solutions for banking, e-commerce, and financial services. Expertise in Python, R, SQL, and advanced libraries such as Scikit-learn, TensorFlow, and PyTorch for effective model development and deployment. Proficient in NLP, deep learning techniques, and generative AI, with a strong focus on real-world applications. Experienced in utilizing cloud platforms and MLOps tools for scalable model management and creating impactful dashboards with Tableau and Power BI.

SDG Chatbot – Qentelli Solutions:

Objective: To develop and deploy an AI-driven chatbot that provides users with quick, accurate, and multilingual information on Sustainable Development Goals (SDGs), enhancing engagement and accessibility.

Technologies Used: Python, Flask, Azure Bot Service, Azure Cognitive Services, Azure Machine Learning Studio, Azure DevOps, LUIS

Responsibilities:

· Developed a Flask-based backend to handle chatbot workflows and user requests.

· Integrated Azure Cognitive Services and LUIS for intent recognition, sentiment analysis, and NLU.

· Deployed the chatbot using Azure Bot Service with CI/CD via Azure DevOps.

· Incorporated multilingual support and real-time feedback mechanisms.

· Used Azure ML Studio to fine-tune models for accurate responses.

· Collaborated with stakeholders to align chatbot capabilities with business goals.

Sanction Screening for AML Compliance – Ernst & Young LLP (Banking Client)

Objective: To build a scalable and intelligent name-matching system to enhance AML compliance by detecting sanctioned entities more effectively while minimizing false positives.

Technologies Used: Python (Pandas, NumPy, Scikit-learn, Matplotlib), MySQL, NLTK, Spacy, TF-IDF, Word2vec, PyTorch, Docker, AWS

Responsibilities:

· Aggregated and pre-processed data from global sanction lists and transactional records.

· Applied NLP techniques such as tokenization, lemmatization, phonetic encoding, and embeddings.

· Built ML models (logistic regression, neural networks) to classify potential matches.

· Engineered features to capture linguistic and semantic similarity.

· Performed hyperparameter tuning and cross-validation.

· Deployed solution via Docker on AWS and documented the system for client use.

Prize Optimization Engine – Freelance Project (E-commerce Client)

Objective: To create a machine learning system for optimizing prize distribution in marketing campaigns to boost customer engagement and maximize promotional ROI.

Technologies Used: Python (Pandas, NumPy, Scikit-learn, Matplotlib, Seaborn), MySQL, PyTorch, Tableau, Jupyter Notebooks, Docker, Kubernetes, AWS

Responsibilities:

· Analyzed past promotional data for engagement trends.

· Developed predictive models (random forests, neural networks) to estimate response rates.

· Applied optimization algorithms like linear programming and genetic algorithms.

· Created Tableau dashboards for performance tracking.

· Deployed using Docker and Kubernetes.

· Implemented retraining pipelines for continuous improvement.