PROFESSIONAL SUMMARY

Seasoned Data Scientist and NLP Engineer with 4+ years of specialized expertise in Large Language Models (LLMs), Natural Language Processing, and Generative AI. Proven track record in architecting and deploying production-grade NLP solutions, fine-tuning transformer models, and building intelligent document processing systems. Adept at leveraging state-of-the-art language models for business-critical applications including conversational AI, intelligent search, and automated content analysis.

PROFESSIONAL EXPERIENCE

• Lead NLP architect for GenAI initiatives, fine-tuning and deploying LLMs (BERT, GPT, T5) and diffusion models for enterprise content generation, achieving 35% improvement in user engagement metrics
• Architected conversational AI chatbot using advanced transformer models and semantic matching algorithms to intelligently recommend 250+ courses based on natural language understanding of candidate profiles and educational backgrounds
• Designed and implemented enterprise RAG (Retrieval-Augmented Generation) search solution using vector databases and embedding models to enable intelligent querying of knowledge bases for investors and customers
• Developed custom prompt engineering frameworks and fine-tuning pipelines for domain-specific LLM applications, reducing inference latency by 40%
• Led NLP-driven business intelligence initiatives that directly supported customer acquisition and engagement strategies through sentiment analysis and intent classification

• Spearheaded AI-Powered Marksheet Verification System: Developed end-to-end NLP pipeline processing 100,000+ academic documents with 94% accuracy using transformer-based classification models
  • Implemented intelligent OCR with named entity recognition (NER) for automatic extraction of student details, grades, and institutional information
  • Built cross-referencing validation system using semantic similarity models to match extracted data against official university records
  • Created automated anomaly detection framework identifying discrepancies and errors in marksheet data
  • Designed natural language query interface for generating custom student result combinations and reports
• Engineered multi-modal document understanding pipeline combining computer vision and NLP for certificate authentication, including text extraction, layout analysis, and security feature detection
• Developed named entity recognition and information extraction systems for structured data extraction from unstructured educational documents
• Implemented face recognition systems with liveness detection using deep learning for enhanced authentication security

• Architected Autonomous Log Analyzer & Incident Predictor: Built offline, open-source AI system for intelligent log analysis and proactive issue prediction
  • Developed log parsing and normalization engine handling both structured (JSON) and unstructured (text) log formats
  • Implemented unsupervised anomaly detection using clustering algorithms and statistical methods for real-time issue identification
  • Integrated local LLM (Llama, GPT4All) for log summarization and natural language explanation generation
  • Created feedback loop mechanism enabling continuous model improvement through user interactions
  • Built intuitive CLI interface with text/HTML output supporting end-to-end workflow: ingestion → detection → explanation → feedback
  • Achieved successful deployment across multiple production environments with demonstrable reduction in incident response time
• Designed and trained transformer-based models for text classification, sentiment analysis, and business forecasting applications
• Built scalable NLP pipelines incorporating tokenization, embedding generation, feature engineering, and model validation
• Deployed production NLP models to AWS and Azure using containerization (Docker) and CI/CD practices, ensuring 99.9% uptime

• Developed sentiment analysis models using BERT and distilBERT for analyzing user feedback across 50,000+ text samples, providing actionable insights for UX improvements
• Built automated text analytics workflows generating comprehensive NLP-driven reports on user engagement patterns and content trends
• Implemented multilingual text extraction systems for ID verification using Tesseract OCR and language detection models
• Applied topic modeling and semantic analysis to educational content, improving content recommendation accuracy by 28%
• Created behavioral analytics framework combining NLP insights with user retention strategies

KEY NLP & LLM PROJECTS

TECHNICAL SKILLS

EDUCATION & CERTIFICATIONS

Copyright Registration (2023) - Programs for Sentiment Analysis and NLP Tasks

KEY ACHIEVEMENTS