AI Agents for Food-Delivery - LangChain, LSTM, FastAPI Fix

We have a production AI agent system for food-delivery operations (1,500+ orders/day). We need an experienced AI developer to debug, optimize, and extend the system. What We Need Fixed: Demand forecasting (LSTM) is drifting - need to retrain with fresh data and improve accuracy Courier dispatch agent occasionally makes suboptimal decisions - need to debug routing logic and add weather/traffic weighting LangChain agent workflow has memory issues in multi-step chains - need to optimize context management with LLM Prompt engineering ETAs are sometimes off - need to retrain delivery time prediction model with Deep Learning (PyTorch) SaaS dashboard (FastAPI + React) is slow on peak load - need performance optimization MLflow tracking is broken - need to fix experiment logging and model registry Automation pipeline for daily retraining fails intermittently - need to debug and stabilize with Python scripting What We're Looking For: AI Agent Development with LangChain and LLM Prompt engineering Machine Learning and Deep Learning (LSTM, PyTorch) for forecasting Python + FastAPI backend with Full-Stack Development (React/TypeScript) MLOps (MLflow, Docker) on AWS Automation pipelines for model retraining and monitoring SaaS Development experience How to Apply: Share your experience debugging production ML systems and optimizing LangChain agents.