Vedant Madane

PROFESSIONAL SUMMARY

Senior NLP Engineer and Generative AI Specialist with 4+ years of production-level experience in developing and deploying Large Language Models, advanced RAG systems, and enterprise-grade generative AI applications. Expert in applying NLP techniques including text classification, summarization, information retrieval, knowledge extraction, and conversational AI using both traditional ML and generative AI approaches. Proven expertise in LLM fine-tuning, VDB optimization, prompt engineering, and scaling AI solutions on cloud infrastructure. Strong background in analyzing large-scale datasets and developing end-to-end ML/DL pipelines from conception to production deployment.

PROFESSIONAL EXPERIENCE

Maharashtra Knowledge Corporation Limited | Aug 2021 - Present

Senior Porject Associate -2 | Generative AI & NLP

Aug 2024 - Present

Lead architect for enterprise-grade generative AI applications, implementing advanced RAG systems with optimized vector databases (Pinecone, ChromaDB) and custom LLM fine-tuning pipelines achieving 35% improvement in user engagement
Developed and deployed production-level conversational AI chatbots using transformer architectures (BERT, GPT, T5) with semantic matching algorithms to intelligently recommend 250+ courses through natural language understanding of candidate profiles
Designed enterprise search solutions leveraging retrieval-augmented generation for knowledge base querying, enabling data-driven insights for investors and stakeholders
Built custom prompt engineering frameworks and implemented LLM evaluation metrics, reducing inference latency by 40% while maintaining output quality
Led cross-functional AI initiatives integrating NLP-driven sentiment analysis and intent classification to support business development and customer acquisition strategies

Senior Project Associate -1 | Worked on NLP & Document Intelligence

Aug 2023 - Jul 2024

Architected AI-Powered Marksheet Verification System: End-to-end NLP solution processing 100,000+ documents with 94% accuracy using transformer-based classification models
- Implemented intelligent OCR with Named Entity Recognition (NER) for automatic extraction and structuring of student data, grades, and institutional information
- Built semantic similarity validation system cross-referencing extracted data against official university records with automated anomaly detection
- Designed natural language query interface for flexible result generation and reporting
Engineered multi-modal document understanding pipeline combining computer vision (LayoutLM) and NLP for certificate authentication, text extraction, and security feature detection
Developed production-ready information extraction systems for structured data parsing from unstructured educational documents at scale
Implemented face recognition and liveness detection systems using deep learning for enhanced authentication security

Project Associate | Worked on Natural Language Understanding (NLU) and Document Extraction

Aug 2022 - Jul 2023

Built Autonomous Log Analyzer & Incident Predictor: 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) formats with robust data cleaning pipeline
- Implemented unsupervised anomaly detection using clustering algorithms and statistical methods for real-time issue identification
- Integrated local LLMs (Llama, GPT4All) for log summarization and natural language explanation generation without external API dependencies
- Created feedback loop mechanism enabling continuous model improvement through user interactions
- Built CLI interface supporting end-to-end workflow: ingestion → detection → explanation → feedback with HTML/text output
Designed and trained transformer-based models (BERT, DistilBERT, RoBERTa) for text classification, question answering, and business forecasting applications
Built scalable NLP pipelines incorporating tokenization, embedding generation (Word2Vec, FastText, ELMO), feature engineering, and hyperparameter optimization
Deployed production NLP models to AWS and Azure cloud infrastructure using Docker containers and CI/CD practices, ensuring 99.9% uptime and efficient resource utilization

Project Trainee | Worked on NLP & Analytics

Aug 2021 - Jul 2022

Developed sentiment analysis and text classification models using BERT and DistilBERT for analyzing 50,000+ user feedback samples, providing actionable insights for UX improvements
Built automated text analytics workflows generating comprehensive NLP-driven reports on user engagement patterns, content trends, and behavioral analysis
Implemented multilingual text extraction systems using Tesseract OCR with language detection models for ID verification applications
Applied topic modeling (LDA) and semantic analysis to educational content, improving recommendation system accuracy by 28%
Developed data scraping architectures to extract and organize information from external sources for generating business insights

KEY PROJECTS

AI-Powered Marksheet Verification System

Enterprise Document Intelligence Solution

Production-grade NLP system for automated academic document verification and validation
Technologies: BERT, LayoutLM, Tesseract OCR, NER, Semantic Similarity, Anomaly Detection
Impact: 90% reduction in manual verification time, 94% accuracy, scalable to 100,000+ documents

Autonomous Log Analyzer & Incident Predictor

Offline Generative AI System for Predictive Maintenance

Open-source AI tool with local LLM integration for intelligent log analysis and proactive issue prediction
Technologies: Local LLMs (Llama/GPT4All), Unsupervised ML, Clustering, Pattern Recognition
Impact: Fully offline solution with real-time anomaly detection and natural language explanations

TECHNICAL SKILLS

Generative AI & LLMs: GPT, BERT, T5, LLaMA, RAG, Vector Databases, Fine-tuning, Prompt Engineering, LLM Evaluation

NLP Techniques: Text Classification, Summarization, Q&A, Information Retrieval, NER, Sentiment Analysis, Topic Modeling

NLP Algorithms: Transformers, Word2Vec, FastText, ELMO, NLU, NLG, Semantic Search

ML/DL Frameworks: PyTorch, TensorFlow, HuggingFace Transformers, spaCy, NLTK, Scikit-learn, LangChain

Programming: Python (Expert), SQL, JavaScript, C++

Data Processing: Pandas, NumPy, Data Scraping, Large-scale Data Analysis

Cloud & DevOps: AWS (SageMaker, S3, Lambda), Azure ML, Docker, FastAPI, Flask, CI/CD

Vector Databases: Pinecone, ChromaDB, FAISS

EDUCATION & CERTIFICATIONS

Bachelor of Engineering - Electronics and Telecommunication

Graduated June 2021

University of Mumbai | CGPI: 7.49

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

KEY ACHIEVEMENTS

4+ years specialized experience in NLP, ML, DL with 1+ year in production-level generative AI application development

Successfully deployed enterprise-grade LLM applications and RAG systems serving 100,000+ users with 99.9% uptime

Deep theoretical and hands-on expertise in NLU, NLP, NLG, and state-of-the-art transformer architectures

Proven ability in developing end-to-end NLP/ML/DL/GenAI solutions with focus on data quality, scalability, and deployment best practices

Experience analyzing large-scale user-generated content and process data in cloud environments