Uday B.

Key Skills

Software

CrowdSource

Data Annotation Tech

Dataloop

Top Subject Matter

Banking customer support intelligence and conversational NLP

Customer support intelligence: ticket categorization and sentiment/urgency NLP

Legal Services & Contract Review

Top Data Types

Text

Document

Top Task Types

Fine-tuning

Entity (NER) Classification

Classification

Freelancer Overview

AI/ML Engineer at Centific (NLP model training and fine-tuning for customer experience). Brings 6+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include AWS SageMaker and Azure ML. Education includes Master of Science, University of North Texas (2024) and Bachelor of Technology, Sphoorthy Engineering College (2022). AI-training focus includes data types such as Text and labeling workflows including Fine-tuning and Entity (NER) Classification.

Labeling Experience

AI/ML Engineer at Centific (NLP model training and fine-tuning for customer experience)

AWS SageMakerTextFine-tuning

Built and deployed transformer-based NLP pipelines to classify customer intent and sentiment from large-scale banking interaction text. Applied parameter-efficient fine-tuning methods (LoRA/PEFT) on proprietary banking datasets while supporting stringent data privacy and governance requirements. Implemented retrieval-augmented conversational flows to generate appropriate responses for account inquiries and financial guidance. • Used AWS SageMaker to fine-tune BERT/RoBERTa models for intent detection and query classification • Employed Hugging Face Accelerate and AWS EC2 with LoRA/PEFT for efficient model adaptation • Implemented real-time conversational pipelines with LangChain and FAISS for relevant retrieval • Tracked and optimized experiments with Optuna and MLflow to improve model accuracy and satisfaction metrics

2024 - Present

AI/ML Engineer at EPAM System (training and deployment of labeled NLP datasets/models)

TextEntity (NER) Classification

Developed and deployed NLP models for automating ticket categorization and sentiment/urgency analysis from customer support text such as tickets and chat logs. Prepared clean, labeled datasets suitable for training classification and NLP models across issue types and customer sentiment. Tuned models using explainability techniques and deployed real-time scoring services integrated into a support platform. • Built ETL pipelines to create labeled training datasets from support tickets, chat logs, and usage feedback • Trained classification models for issue type categorization (hardware, software, billing) and sentiment-based urgency prediction • Used Grid Search and SHAP to tune models and improve transparency of predictions • Deployed models via Azure ML with Docker/Kubernetes and scheduled monthly retraining with Airflow based on new data

2021 - 2023

Education

Work History