M
Mehul Gilotra
About
Detail
Karnataka, India
Résumé
Jobs
verified_user
0% verified
Software DeveloperOracleAug 2023 - Current (3 years)Developed an advanced agentic AI chatbot with LangGraph and domain-specific RAG architecture, achieving 90%+ query accuracy and reducing manual support overhead. Integrated Oracle Digital Assistant (ODA) with custom NLP models for document parsing, translation, and entity recognition, enhancing contextual understanding in enterprise tools. Built and deployed a vector database-backed RAG pipeline for unstructured data ingestion and semantic search, improving retrieval latency by 35% and enabling scalable ML-driven knowledge access. Engineered AI-driven automation pipelines with LangChain and FastAPI, enabling LLMs to autonomously handle real-time queries and trigger OCI workflows without human intervention. Created a conversational HR assist- EInternErnst & Young (EY)May 2023 - Jun 2023 (2 months)Automated backend pipelines in Python for insurance reporting, reducing manual work by 70% and ensuring reproducible audit runs.
- Research InternIndian Institute of Science (IISc)Aug 2021 - Jan 2022 (6 months)Engineered multilingual NLP for low-resource languages, improving model accuracy by 25% and enabling scalable backend integration of ML models.
Education
verified_user
0% verified
- OOracle Cloud Infrastructure (OCI)Aug 2023 - Current (3 years)
- PPythonAug 2023 - Current (3 years)IIT Madras
- MMachine LearningAug 2020 - Jan 2021 (6 months)Duke University (Coursera)
- RB.E., Computer Science & EngineeringRV College of Engineering, BengaluruJan 2019 - Jan 2023 (4 years 1 month)
Projects (professional or personal)
verified_user
0% verified
- OOracle Vector Search CertificationAug 2023 - Current (3 years)
- CCustomer Churn PredictionMay 2023 - Jun 2023 (2 months)Built an end-to-end ML pipeline using TensorFlow, Scikit-learn; achieved 92% accuracy and improved recall by 20% through hyperparameter tuning and feature engineering.
- SSign Language TranslatorFeb 2023 - Mar 2023 (2 months)Developed a real-time gesture interpreter and deployed it as a web service via API, achieving 91%+ accuracy with speech output for accessibility.