Farhad Davaripour, PhD.

As an MLOps developer, simulation specialist, and research engineer with 7 years of experience, I specialize in creating innovative data-driven solutions. My expertise includes leveraging machine learning, generative AI, and synthetic data from finite element simulations to analyze stress in steel pipelines and tackle complex industry challenges. Currently, at Arcurve Inc., I utilize cloud services (e.g., Azure, AWS) and analytical platforms (e.g., Databricks, Snowflake) to design and develop end-to-end machine-learning and ETL pipelines, as well as GenAI-based applications, to address critical industry needs. 

Webinar Series

Arcurve Webinar Series Ep. 1: Research Agent - An Application of Agentic Workflow

Arcurve Webinar Series Ep.2: Deep Dive into the Research Agent Leveraging AI Agentic Workflows

Arcurve Webinar Series Ep.3: Bridging AI and Engineering Standards - Design & Compliance Handling

Arcurve Webinar Series Ep.4: Augmenting Simulation Engineering with AI Agents: Integrating Finite Element Analysis With Agentic AI

Educational Contents/Apps

1. Linear Regression Core Principles

In this lecture you will learn how to build a linear regression model from scratch.

GitHub page: Link

Run demo notebook on Google Colab: Link

2. Next Token Prediction

In this lecture, you will learn how Temperature and Top-P control the randomness in LLM-generated outputs 

Dashboard: Link

3. Intro to Prompt Engineering

In this lecture, you will learn how to implement the most popular prompt engineering techniques.

GitHub page: Link

Run demo notebook on Google Colab: Link

4. Intro to Retrieval Augmented Generation (RAG)

In this lecture, you will learn how to implement a basic RAG pipeline.

GitHub page: Link

Run demo notebook on Google Colab: Link

5. Intro to LLM-based Function Calling

In this lecture, you will learn how to implement a basic function calling workflow.

GitHub page: Link

Run demo notebook on Google Colab: Link

6. ReAct Agentic Implementation

This lectures goes through an implementation of a ReAct Agentic Workflow from scratch.

GitHub page: Link

App: Link

7. AI in Pipeline Engineering

This lectures goes through an end to end ML workflow to train a pipeline ILI data analyzer.

GitHub page: Link

Colab Notebook: Link

App: Link

8. GenAI in Pipeline Engineering

This lectures goes through a development of a ReAct Agentic Workflow to parse and execute ALA 2005 design guideline for buried steel pipeliens.

GitHub page: Link

App: Link

9. AI Assisted FEA

This lecture explores implementing Abaqus (a Finite Element Analysis tool) as part of a ReAct agent powered by OpenAI's LLM, automating finite element simulations, including input generation, job execution, stress analysis, parametric studies, and sensitivity analysis, streamlining workflows for improved decision-making.

GitHub page: Link

Distilled Notes

Stanford-CS229-Spring2023

CS229 course notes from Stanford University on machine learning, covering fundamental concepts and algorithms.

GitHub page: Link

App: Link

Educational Contents for Pipeline Engineers

AI in Pipeline Engineering

This course explores various machine learning (ML) implementations specifically targeting preventative maintenance in pipeline engineering challenges. 

GitHub page: Link

GenAI in Pipeline Engineering

This course explores various advanced GenAI (e.g., ReAct Agentic Workflows) implementations in the field of pipeline engineering.

GitHub page: Link

Applications

           DocsGPT

This app allows users to easily query documents with varied formats using OpenAI's GPT-3.5 language model.

GitHub page: Link

Tutorial: Link

Google Colab: Link

CFRP Reinforced Pipe Bend

This project employs machine learning and synthetic dataset to predict the peak equivalent stress imposed on a CFRP-wrapped HDD overbend.

GitHub Page: Link