Uddagiri Manoj

AI Engineer focused on architecting and developing enterprise-grade AI platforms using Large Language Models (LLMs), agentic AI, Retrieval-Augmented Generation (RAG), and distributed inference systems. Experienced in designing scalable multi-tenant AI infrastructure with intelligent model routing, long-term memory, vector search, secure tenant isolation, workflow orchestration, and production-ready AI services. Skilled in GPU optimization, vLLM, Qwen models, LangGraph, smolagents, Qdrant, PostgreSQL, MinIO,

• Developed and tuned demand forecasting models using ARIMA, Prophet, and LSTM to improve logistics prediction accuracy. • Applied optimization algorithms using OR-Tools to reduce delivery time and operational logistics costs. • Performed data cleaning, transformation, analysis, and visualization using Pandas, NumPy, and Matplotlib. • Supported end-to-end integration of machine learning models into the CADDIE production logistics platform. • Collaborated with product and engineering teams to align machine

1. Built and tuned demand forecasting models (ARIMA, Prophet, LSTM) supporting logistics forecasting and route optimization for the CADDIE platform. 1. Applied OR-Tools optimization to delivery routing, contributing to measurable reductions in delivery time and logistics cost. 1. Collaborated with product teams to integrate ML models, built in Python (Pandas, NumPy, Matplotlib), into a live platform.

Data classifying for satellite images using the Xception modelinvolves leveraging a deep convolutional neural network architecture. Xception is particularly well-suited for this task due to its ability to capture intricate patterns in high-dimensional data. By training Xception on labeled satelliteimage datasets, the model can effectively classify imagesinto various categories.