Shayan Sadigh

Location:Santa Clara, CA
Email:ssadigh.contact@gmail.com
Github:https://github.com/shayanpersonal (150+ stars)

Summary

• Prompt Engineer | Ex-Amazon AI
• (2018) Won 10th in $1.5 million Kaggle competition - trained custom MVCNN-RNN neural net (Github)
• (2018) Proposed new deep learning layer - "checkered subsampling" (arXiv)
• I've been working on AI since 2016, training models across modalities (vision, language, audio) and studying papers across various topics (CNNs, RNNs, LLMs, learning rules, reinforcement learning, etc).

Work Experience

aurascape.ai

AI Researcher (Feb. 2024 - Current)

• Building transformer-based AI agents. Founding team.

Santa Barbara City College (SBCC)

Adjunct Professor (Aug. 2023 - Dec. 2023)

• Walked students through various aspects of designing and training deep learning models. Discussed linear layers, nonlinearities, normalization, convolution, aspects of transformers such as attention layers and embeddings. Discussed backpropagation and loss functions.
• Implemented k-nearest-neighbors and neural nets including a "small language model" in PyTorch for homework. Trained models on MNIST and tiny-shakespeare. Used MoeBot to generate curated study material.
• Taught CS114: Intermediate Python. SBCC is frequently ranked among the top five city colleges in the United States.

moebot.ai

Prompt Engineer (2021-2023)

• Owner of moebot.ai. Built subscription-based LLM agent for Discord servers. Makes discrete decisions such as deciding whether to respond or send emoji reactions via prompt engineering. Stores and recalls rudimentary "memory" of users (a form of RAG or Retrieval Augmented Generation). Runs on Google Cloud, Redis, Flask, Python. Generates revenue through Stripe transactions.
• Volunteered to speak at Santa Barbara City College and later taught as a professor. Used MoeBot as an AI teaching assistant.
• Built various smaller projects.

Amazon Web Services (AWS)

Deep Learning Engineer II (Mar. 2020 - Dec. 2020)

• Helped integrate Apache TVM deep learning compiler into Amazon SageMaker's PyTorch fork.

Palo Alto Networks

Senior Software Engineer (Oct. 2019 - Mar. 2020)
Software Engineer (Sep. 2018 - Sep. 2019)

• Awards: Project of the Quarter - Machine Learning Antivirus
• Proposed and implemented new stack-based pattern matching algorithm which improved threat coverage in PAN-OS for a certain class of threats, written in C.
• Helped integrate machine learning models into resource-constrained firewall.
• Solved dozens of the team's open tickets.
• Brought up and helped train two junior engineers.

Other

Spring 2018 | Teaching Assistant @UCSB (CS176A: Intro to Computer Networks)
Summer 2017 | Software Engineer Internship @Palo Alto Networks
Summer 2016 | Software Engineer Internship @Thermo Fisher Scientific
2015 | Web developer @UCSB Brain Initiative

Projects

• stacked-autoencoder-pytorch (100+ stars) - a popular PyTorch autoencoder.
• Checkered CNN (20+ stars) - my deep learning technique, checkered subsampling.
• Passenger Screening Model (20+ stars) - solution to $1.5 million passenger screening challenge.
• GauchoMap Chrome Extension - 200+ users when I was maintaining it (2016).

Education

• 5-year accelerated B.S. + M.S. program @ University of California, Santa Barbara
• Master of Computer Science (2018)
• Bachelor of Computer Science (2017)

Miscellaneous

• Top-30(?) Dance Dance Revolution player in California.
• Ranked 1st globally (tied) in popular competitive arena shooter.
• 2nd-degree black belt in Soo Bahk Do.
• Video game soundtrack composer.
• First-generation college graduate.
• Languages: English (native), Farsi (native, but very rusty)