Name: 🤖 [Special Event!] Federal Civilian Contractor AI Meetup 🤖
Start: 2024-08-01T17:30:00-04:00
End: 2024-08-01T20:00:00-04:00
Location: Pyramid Systems Inc

Make history and join the nascent AI Meetup Community!

Augusts Topic: Hot New AI Topics

Connect with peers, share insights, and broaden your professional network.

🍕Join us for an evening of learning, discussion and free Pizza 🍕

**Who Should Attend:**

* Federal Civilian Contractors
* AI Enthusiasts, Data Scientists and Thought Leaders
* Government Experts
* Anyone interested in the intersection of AI and Federal Civilian Contracts

**Speakers:**

* Dr. Abigail Haddad - Data Scientist and Author
* Mona Mona - AI/ML (Artificial Intelligence Machine learning) specialist at Google Public Sector and Author

Agenda:

* 5:30 - 6:15 pm - Meet, Greet and Pizza
* 6:30 - 8:00 pm - Speakers

John Sobanski

Washington Federal Civilian AI Meetup Group

pyramid

Technology

Professional Networking

Machine Learning

Artificial Intelligence

Artificial Intelligence Applications

Government Contractors

Government Contracting

AI Algorithms

* Speaker 1: Dr. Abigail Haddad

Dr. Abigail Haddad is a data scientist who was working on automating LLM evaluations for whether they can assist in tasks like building bioweapons. Previously, she worked on research and data science for the Department of Defense, including at the RAND Corporation and as a Department of the Army civilian. Her hobbies include analyzing federal job listings and co-organizing Data Science DC. She blogs at[ Present of Coding](https://presentofcoding.substack.com/).

* Abstract

**Evaluating Large Language Model Risks: Aligning Testing Strategies with Threat Models**

As new, more advanced Large Language Models (LLMs) are developed, both companies and government agencies face challenging questions about how to approach pre-release testing to ensure these models cannot be misused for certain purposes. Concerns span a range of areas. However, the appropriate testing strategy depends heavily on the anticipated threat model.
This talk explores how different threat scenarios - from the continued discovery of easily-exploitable jailbreaks to the potential theft of model weights - dictate whether black-box jailbreaking tests or more comprehensive white-box capabilities assessments are necessary. On one hand, if jailbreaking vulnerabilities are readily identified and patched and model weights remain secure, then a focus on black box testing is sufficient. But if model weights could be stolen and misused by sophisticated actors, a deeper evaluation of the underlying model capabilities becomes critical.
By outlining these varying threat models, the presentation illustrates how testing methodology can be tailored to the most likely risks. The talk concludes by suggesting areas for further research to better understand the real-world dynamics of LLM security, which can inform appropriate testing methodologies for companies like OpenAI and Anthropic, as well as government agencies like NIST tasked with advancing AI safety.

* **Speaker 2: Mona Mona**
    Mona currently works as an AI/ML (Artificial Intelligence Machine learning) specialist in Google Public Sector. She was a Sr AI/ML specialist Solutions Architect at Amazon before joining Google. She has earned her masters in Computer Information Systems from Georgia State University.
    Mona Mona is a published author of books [Natural Language Processing with AWS AI Services: Derive strategic insights from unstructured data with Amazon Textract and Amazon Comprehend](https://www.amazon.com/Natural-Language-Processing-AWS-Services-ebook/dp/B09DDHQDB6) and Google Cloud Certified Professional Machine Learning Study Guide. She has authored 19 blogs on AI/ML and cloud technology and a co-author on a research paper on [CORD19 Neural Search](https://www.amazon.science/publications/aws-cord-19-search-a-neural-search-engine-for-covid-19-literature) which won an award for Best Research Paper at the prestigious [AAAI (Association for the Advancement of Artificial Intelligence)](https://aaai.org/) conference.
    Mona Mona is a member of [SWE(Society for women engineers)](https://swe.org/) and [ACM(The Association for Computing Machinery)](https://www.acm.org/) and she mentors early career graduates looking to pursue careers in AI and Cloud Computing.
    She is also a frequent speaker at multiple conferences such as ISMB 2022, AWS Re:Invent 2020,2019,2018 and AWS DC Summit
* **Abstract**
    Mona will discuss developing Generative AI applications on the Google Cloud Platform (GCP). She will address the integration of AI technologies with GCP’s cloud infrastructure. Attendees will learn how to use GCP's suite of AI tools, including Vertex AI for model training and deployment, AutoML for building custom models with minimal coding, and pre-trained models available through the AI Hub. The session covers best practices for data management, scalability, and security in AI application development. Real-world case studies highlight the use cases of Gen AI on GCP, and provides practical roadmap from concept to deployment.

Dr. Abigail Haddad & Mona Mona

🤖 [Special Event!]  Federal Civilian Contractor AI Meetup 🤖

Pyramid Systems Inc

Share

Washington Federal Civilian AI Meetup Group

🤖 [Special Event!] Federal Civilian Contractor AI Meetup 🤖

Washington Federal Civilian AI Meetup Group

Details

Members are also interested in