Linkedin

Vishnu priya P

VIT Vellore

IoT Real-Time Sensor Diagnostic Agent

This project is a 6-node AI agent built with LangGraph that automates IoT maintenance. It processes real-time sensor streams using a hybrid approach: first applying four statistical methods (Z-Score, IQR, Trend, Rate-of-Change) to filter noise, then triggering GPT-4 to analyze confirmed anomalies for root cause diagnosis. Why it matters: Traditional monitoring only flags that something is wrong; this agent explains why it’s wrong and how to fix it. It eliminates false positives and empowers engineers with actionable, autonomous insights rather than raw data alerts.

• Programming Languages: C, C++, Python, Embedded C, Java (basics), SCALA (Basics), VHDL (Basics), MATLAB, Open MP, Linux (Basics). • Development Tools: Power BI, Visual Studio Community, VS code, Code composer studio, MATLAB, MS Office, Node RED, Google Colab, Azure, SQL Server Management Studio. • Frameworks and Technologies: supervised, unsupervised learning, and Deep learning models like CNN, RNN, LSTM BiLSTM, LLMs and transformers like BERT, MLOPS, hugging face, PEFT (e.g., LoRA). • Angular, .NET (C# Basics), REST APIs • Proficient in Data science - Pandas, Numpy, Scikit learn. SQLPlus, PySpark, Joblib, SCP, PyTorch, GAN, RAG, Kafka, Flask, Streamlit, Docker, Gen AI, Object Detection and Image processing. • Embedded: Microcontrollers – NRF Connect SDK, MSP430, Arduino UNO, Raspberry Pi, Edge devices, RTOS, ESP 8266 worked with many wired sensors and motors like Temperature, servomotors etc. • Soft skills: Teamwork, Problem solving, Leadership, Adaptability.

1) Clone the repo git clone https://github.com/YOUR_USERNAME/iot-sensor-agent.git cd iot-sensor-agent # 2) Create .env with your own OpenAI API key cp .env.example .env # then edit .env and set OPENAI_API_KEY # 3) Build image docker build -t iot-agent:latest . # 4) Run batch demo (3 scenarios, simulated sensor data) docker run --rm --env-file .env iot-agent:latest

Yes

Yes

Yes

Yes

Yes, I’m interested