Hello, I'm
Tannu Tiwari
Building scalable multi-agent systems, deep learning models, and intuitive applications. Passionate about turning data into actionable intelligence.
Path Way
Data Analyst Intern
January 2026 – April 2026Ernst & Young, Gurugram, Haryana
- Processed and validated large-scale examination datasets (100K+ records) using SQL (SSMS), improving data accuracy and reducing inconsistencies in result workflows by 57%.
- Automated score validation and reporting using SQL queries, reducing manual effort and improving result processing efficiency by 38%.
AI Intern (Remote)
June 2025 – August 2025HCLTech, Noida
- Developed a scalable multi-agent LLM chatbot using LangChain, RAG, and ChromaDB.
- Built ingestion pipelines (chunking, embeddings) improving response relevance by 40%.
- Integrated autonomous agents for code review and automation, increasing throughput by 35%.
Bachelor of Engineering (CSE)
Sep 2022 – PresentBennett University, Greater Noida
- BTech Computer Science and Engineering specializing in Data Science
Senior Secondary
March 2021 – May 2022Delhi Public School Ghaziabad, Merrut Road
Technical Arsenal
AI/ML & Deep Learning
Tools & Frameworks
Programming & Dev
Research
DeepFake Detection Using Neural Networks
Research PaperDesigned a novel ensemble model combining CNN, ResNet, and Vision Transformer (ViT). Achieved 98.81% test accuracy through fine-tuning and data augmentation.
Featured Projects
DevOps MCP Server
May 2026FastAPI service exposing DevOps workflows as authenticated REST endpoints and MCP tools for AI agents and CI/CD.
Inventory Optimization Platform
March 2025End-to-end demand forecasting system using LSTM, XGBoost, ARIMA, and Prophet. Reduced forecasting error by 25%.
Music Genre Classification System
January 2025Audio ML pipeline utilizing MFCC, mel-spectrogram, and ensemble classifiers. Achieved 90%+ track reliability.
HandWave Harmony
September 2024Gesture-controlled audio system. Decreased motion-to-audio latency by 20ms using MediaPipe and OpenCV.
VisualVerse
March 2024Generates descriptive captions for images by combining MobileNetV2 for feature extraction and LSTM for text generation.