Hi there! 👋 I'm Harsh Mehta
🎯 AI/ML Engineer
AI/ML Engineer with a Master's in Information Science from UW-Madison and 5+ years of experience across machine learning, llm's, ai agents, business intelligence, and enterprise analytics. I build production AI systems — RAG pipelines, document intelligence, voice-to-data agents — and translate complex business requirements into data-driven solutions across healthcare, finance, and enterprise domains.
📚 Education
M.S. in Information Science (Data, ML, Cloud) — University of Wisconsin-Madison (GPA: 3.95/4.0)
Bachelor of Management Studies — University of Mumbai
🔧 Certifications
Databricks Data Engineer Associate | Snowflake SnowPro Core | Azure Data Scientist (DP-100) | Microsoft Certified Power BI Data Analyst (PL-300)
🌐 Let's Connect!
Always open to collaborating on AI/ML, data engineering, or analytics projects.
Pinned Loading
-
Arduino-Starter-ESP32-Echokit---C-code-Starters
Arduino-Starter-ESP32-Echokit---C-code-Starters PublicA beginner-to-intermediate Arduino project showcasing how Arduino abstracts C++ through .ino files while also providing an equivalent pure C++ implementation. Includes LED control using millis(), b…
TypeScript
-
Fieldbox-Sensor-EDA
Fieldbox-Sensor-EDA PublicAutomated EDA pipeline for IoT sensor data: quality checks, anomaly detection, visualizations, and plain-English narrative reports. Shell-based orchestration with file watcher daemon. Live interact…
TypeScript
-
InsightTube-AI
InsightTube-AI PublicAI-driven video intelligence platform with semantic search, distributed data pipelines, and containerized analytics services.
TypeScript
-
Kafka-Login-Event-Stream-Pipeline
Kafka-Login-Event-Stream-Pipeline PublicReal-time Kafka pipeline that routes user login events by device platform (iOS/Android) with missing data handling — Docker, Confluent Kafka, Python.
TypeScript
-
Pytorch_Cnn_Demucs_Audio-cladssification-pipeline
Pytorch_Cnn_Demucs_Audio-cladssification-pipeline PublicAn end-to-end PyTorch pipeline bridging professional audio engineering and deep learning. Features Demucs source separation, custom CNN classification, and 3D TensorBoard latent space projections.
Jupyter Notebook
-
Triple-Ocr-Rag
Triple-Ocr-Rag Public3 production RAG pipelines for mortgage PDFs — DocTR + DeepSeek, Chandra OCR, and 5-tier local fallback. FAISS vector search, BGE/MPNet embeddings, Gradio UI with source citations.
Jupyter Notebook
If the problem persists, check the GitHub status page or contact support.



