Machine Learning & NLP Engineer

🚀 About Me

I am a passionate Machine Learning Engineer with 4 years of experience in developing innovative solutions using AI, predictive modeling, and natural language processing (NLP). Throughout my career, I have worked on a variety of projects, including automated ticket routing systems, intelligent voice assistants, and more, helping organizations optimize their processes and improve user experiences.

Currently, I am focused on gaining expertise in Large Language Models (LLMs), exploring their potential to revolutionize AI applications in areas like natural language understanding, generation, and dialogue systems. My goal is to deepen my knowledge of transformer models and other cutting-edge techniques, while building scalable, real-world solutions on cloud platforms like AWS.

I thrive in collaborative, fast-paced environments and am always eager to learn and experiment with new technologies. Feel free to explore my portfolio to see how I have leveraged machine learning to drive impactful results.

My key areas of expertise include:

Deep Learning (using TensorFlow, PyTorch)
NLP (using Hugging Face Transformers, BERT, GPT-3,LLMs)
Machine Learning (scikit-learn, XGBoost)
Data Preprocessing & Feature Engineering
Model Deployment (AWS, Docker)
Data Structures and Algorithms

💼 Experience

Senior Engineer| NLP | Samsung Research Institute-Bangalore (SRI-B)

December 2021 - July 2023

Developed new framework for the games promotion and better user experience in the Conversational Voice assistant Bixby.
Enhanced data pipelines for preprocessing and feature extraction and were able to acheive 5% increase in better user experience.
Implented POCs for ML and DL research enhancements in the existing architecture of the chat assistant such as integration of LLMs, sentiment enhancement techniques.
Worked with state-of-the-art frameworks like PyTorch, TensorFlow, and Hugging Face Transformers as part of research ideation.
Increased the code quality and maintained the sloc metrics of 95% and above.
Collaborated with cross-functional teams such as CC,ASR,TTS and Linguistics to integrate Bixbychat models into production and optimize performance.

ML Engineer | Wipro Technologies

June 2019 - November 2021

Built predictive models for ticket management and automated routing, developing and presenting the POC.
Worked on NLP tasks, including Named Entity Recognition (NER) and topic modeling.
Implemented data visualization solutions using Matplotlib, Seaborn, and Plotly.
Automated report generation scripts to identify and list backlogs or prioriy tickets, improving process efficiency.

⚙️ Skills

Programming Languages: Python, R, SQL
Libraries & Frameworks:
- Deep Learning: TensorFlow, PyTorch, Keras, Hugging Face, FastAPI
- Machine Learning: scikit-learn, XGBoost, LightGBM
- NLP: BERT, GPT-3, spaCy, NLTK, TextBlob
Tools & Platforms:
- AWS, Docker, Kubernetes, Git, Jupyter Notebooks, Visual Studio Code,Perforce
Data Visualization: Matplotlib, Seaborn, Plotly, Tableau
Model Deployment: Flask, FastAPI, Docker, Kubernetes
Databases: MySQL, MongoDB, Neo4j

💻 Selected Projects

1. Chain of Thought (CoT) in LLMs – Paper Implementation

This project replicates the work on Chain of Thought (CoT) reasoning in Large Language Models (LLMs). The implementation of CoT allows the LLMs to perform better on complex reasoning tasks without requiring multi-shot learning.

Technologies: Python, Hugging Face Transformers, Deep Learning
Features:
- Implemented zero-shot CoT reasoning for LLMs
- Demonstrated improvements in model scaling and performance across various datasets
- Replicated results from the “Large Language Models are Zero-Shot Reasoners” paper by leveraging chain-of-thought reasoning

2. Sentiment Analysis using BERT

A sentiment analysis model built with BERT (Bidirectional Encoder Representations from Transformers) for classifying text (e.g., product reviews) as either positive or negative. This project demonstrates fine-tuning transformer-based models for NLP tasks.

Technologies: Python, TensorFlow, Hugging Face Transformers
Features:
- Fine-tuned a pre-trained BERT model on a custom sentiment dataset
- Used Hugging Face Transformers for efficient model training and deployment
- Deployed the model via a Flask API for easy integration with web apps

3. Machine Translation using Autoencoder and Decoder

A sequence-to-sequence model for machine translation using an autoencoder architecture with an encoder-decoder approach for translating text from one language to another.

Technologies: Python, TensorFlow, Keras, NLP, Neural Networks
Features:
- Built an encoder-decoder architecture for sequence translation
- Trained on parallel corpora for multilingual text translation
- Implemented attention mechanism for improved performance in longer sequences

4. Accessing Credit Risk using Machine Learning

A machine learning model that predicts whether a customer’s profile is risky or safe based on financial data. The model evaluates risk factors such as credit history, loan amount, and payment history.

Technologies: Python, scikit-learn, pandas, Machine Learning
Features:
- Utilized classification algorithms (e.g., Logistic Regression, Random Forest) for risk classification
- Trained and validated on a real-world financial dataset
- Deployed the model as a web app (Link to the deployed version: Credit Risk Web App)

🧠 Education

Master’s in Computer Science (Focus on Data Science & AI/ML)
University of Central Missouri | 2025
Bachelor’s in Information Technology
Jawaharlal Nehru Technological University Anantapur | 2019

📝 Certifications

Deep Learning Specialization (Coursera - Andrew Ng)
Applied Data Science and Machine Learning with Python Excellence (Coding Ninjas)
Natural Language Processing (Udacity)
Data Structure and Algorithms Excellence (Coding Ninjas)

🌐 Connect with me

Feel free to reach out to me for collaborations, job opportunities, or any AI-related discussions!