We’re announcing this meetup late, but it’s going to be incredible!

Our April meetup will focus on “Build Your Own GPT: Open Source GenAI for the Enterprise”, a topic there was considerable interest in at the AI conference we hosted last month in Durham.

Host location will be the Bandwidth Raleigh Office, which is a brand new state-of-the-art facility housing the Bandwidth company. We’ll be in their training rooms, which are extremely nice. *We hosted a meetup here last October and our community LOVED it!

Note, doors open at 6:00 pm for networking, food and beverages. Live programming begins at 6:30 sharp with announcements and thank you’s.

IMPORTANT – ONLY 100 RSVPs ALLOWED
Because we’re using the training rooms, RSVPs can’t exceed 100. So, please hurry as RSVPs will be cut off after the first 100. We apologize in advance for this.

Bring photo ID: All visitors will check in at the front desk and name badge will be printed. All attendees will be walked back to the meetup area.

Venue Address:
Bandwidth Raleigh Office
2230 Bandmate Way
Raleigh, NC 27607
Additional Venue Details, including parking

Special Swag
We plan to make a number of AI conference shirts available at this event, in addition to a number of ATO and We Love Open Source shirts also.

Speaker:
Mark Hinkle – CEO and Co-Founder, Peripety Labs

Bio:
Mark Hinkle has over 30 years of experience in technology, open source, and enterprise software. He served in executive roles at Citrix and The Linux Foundation, spearheading innovative programs in cloud computing, DevOps, and early internet usage. Over the past two years, he has trained thousands of professionals on how to use AI for real-world impact—ranging from beginners to C-suite executives.

Mark publishes [TheAIE.net](http://theaie.net/), a newsletter network with more than 250,000 subscribers dedicated to working smarter with AI. He also founded and organizes the RDU AI Group, which has grown to more than 1,800 members. Alongside his training for billion-dollar enterprises seeking to adopt AI strategies, Mark is frequently engaged by corporations looking to upskill their workforce, leveraging his deep expertise in early technology adoption to deliver actionable results.

Talk Title:
Build Your Own GPT: Open Source GenAI for the Enterprise

Abstract:
The rapid rise of generative AI has created big opportunities—and even bigger dependencies. This session is for those who want to build an “Enterprise GPT” capabilities on their own terms, using open source software that’s proven, extensible, and production-ready.

We’ll break down the core components required to build an open source GenAI stack that works inside real organizations:

• Open weight foundation models (e.g., LLaMA, Granite, Mistral, IBM Granite, DeepSeek)
• Inference engines that can actually scale (e.g., vLLM, TGI)
• Vector databases built for retrieval-augmented generation (e.g., FAISS, Weaviate, Milvus, Chroma)
• Orchestration frameworks (e.g., LangChain, CrewAI, IBM Bee, Obot) that tie it all together

We’ll talk through what it takes to get GenAI into production—securely, cost-effectively, and without locking yourself into proprietary infrastructure. Expect straightforward advice on managing context windows, test-time compute, and system governance.

Bonus Session:
After the talk, we’ll walk through real-world prompt engineering strategies. We’ll cover prompt patterns that work with open source models, how to improve model outputs without touching the weights, and how to design effective zero-shot, few-shot, and other prompts. And test them with the open source models via Hugging Face. This segment will give you tactical ways to improve output quality for your enterprise use cases.

If you want to run GenAI inside the firewall—or just want more control over how your AI systems behave—this session gives you the map and the tools to build it.

Questions?
Contact us at info@allthingsopen.org.