Held virtually from February 24–28, Intel® Liftoff Days brought together eleven AI startups for a focused acceleration sprint. With workshops, 1:1 mentoring, and product demo sessions, the week was designed to help teams move faster, test more ideas, and get direct support from Intel’s AI team.
Intel® Liftoff Days is part workshop, part development sprint and fully focused on building better AI products.
Our Goals for the Event
Intel® Liftoff Days was designed for progress. Over one week, participating startups got hands-on with Intel’s AI infrastructure, including the Intel® Tiber™ AI Cloud, and applied those tools to real product development goals.
Startups came in with clear objectives, from improving multi-agent reasoning to hosting live drone video and had five days to build, test, and refine with help from Intel engineers.
The Challenges
Our startups got hands-on with advanced AI topics such as Zero-shot LLM inference, RAG pipelines, Traditional ML, MAAS, Computer Vision, and Image Generation.
Judging was based on progress made during the week, technical depth, clarity of demo, and practical application. Extended access to Intel’s tools and cloud resources was awarded to top performers.
The agenda included multiple workshops designed to equip startups with the tools needed to optimize their projects:
Get Started with Intel® Tiber™ AI Cloud: Hands-on walkthrough of setting up environments, launching models, and optimizing workflows.
Model-as-a-Service (MaaS): How startups can use LLM APIs (like Llama 3.1 or Qwen 2.5) without running models locally.
OPEA: An open-source framework for deploying enterprise GenAI solutions—built for interoperability and scale.
LLM API Plug-and-Play: Code samples to integrate chatbots and RAG workflows using Intel-compatible APIs.
Synthetic Data Generation:Training and fine-tuning smaller LLMs using synthetic datasets built on vLLM.
Agent and Chatbot Prototyping: Quick-start templates for OpenAI, DeepSeek, and ITAC APIs.
Selling AI Products: A practical workshop for early-stage founders on how to sell what you’ve built.
Participating startups
Data Centric MLOps SaaS application for enterprise ML teams to manage NER projects.
|
Applied AI research lab with a focus on embedding models for niche use cases including low-represented languages - document embedding models for document-image RAG. |
Explainable AI solutions tailored to detect duplications and contradictions within legal documents and datasets with unparalleled accuracy and efficiency. |
A universal and semantic focused low resource language platform. Using a no code user interface, communities and individuals are able to finetune these scalable language models with their data corpus in a secure, private (encrypted) environment that is cost effective. |
Developing quantum computing using photonics technology, focusing on the relativistic properties of photons to improve efficiency. The company uses these properties to support the miniaturization of quantum computing devices and to recycle heat and energy losses. |
Autonomous drone software for first responders, assisting in search-and-rescue missions with real-time data and autonomous decision-making support. |
Reama Chat, a conversational chatbot designed for renewable energy technicians to improve efficiency in field operations |
AI-powered shopping assistant for e-commerce platforms helping users by recommending products based on browsing behavior, managing customer interactions & driving sales through smart suggestions.
|
Agentic AI platform for creating intelligent chat agents. Helps businesses build customizable agents for customer service, sales, and engagement. |
Web based app to help scientists improve the manufacturing of cell therapeutics.
|
AI based chatbots & agents for education, legal, retail, and advertising. |
|
Projects We Loved
Parawave
Parawave came into the week with a clear goal: migrate their AI models to Intel’s cloud infrastructure and enable low-latency video sharing for first responders using their app, Shepherd.AI. By the end of the sprint, they had achieved just that - connecting their live-streaming server, running on Intel Tiber™AI Cloud, to their mobile and web platforms. Their work helps emergency teams get smarter insights from drone footage, faster. In recognition of their progress and the potential of their mission, Parawave was awarded one month of free access to vision-language models on Intel Tiber™ AI Cloud.
Reama AI
Reama AI had already built their core product on Intel Tiber™ AI Cloud and used Intel Liftoff Days to take it further - focusing on faster LLM reasoning and adding multi-agent capabilities for better customer interactions. Throughout the week, the team enhanced their product’s speed and functionality, making strong use of Intel’s model-as-a-service infrastructure. Their commitment to building a viable and responsive tool for real-world use earned them two weeks of beta access to LLM APIs.
The Mango Jelly
The Mango Jelly team hit the ground running with fast integration of Intel’s API services into their platform. Their goal was to plug Intel’s capabilities directly into their generative workflows and they made it look easy. With a strong foundation and clear product vision, they demonstrated immediate value from the tools provided. For their swift progress and thoughtful execution, they also received two weeks of free API access.
Pixel ML + Parawave
Pixel ML entered the week curious to see how far they could push agentic reasoning using the DeepSeek R1 model. In an unexpected but exciting collaboration with Parawave, they began testing new approaches that could bring reasoning models into real-time drone applications. This cross-team effort sparked new ideas and showed real promise—we’re looking forward to continuing this conversation and sharing their learnings in an upcoming co-publication.
Lessons We Learned
Startups don’t need long timelines to make real progress. With the right tools and the right feedback, five days is enough to go from concept to working demo.
We also saw that small changes, like easier access to infrastructure or API compatibility, can unlock big steps forward, especially when startups are experimenting and iterating fast.
Breakthroughs on Intel®
Startups used Intel® Tiber™ AI Cloud to deploy models, generate synthetic data, and run inference at scale. The plug-and-play APIs helped teams skip setup time and focus on testing what mattered.
If you're part of the Intel Liftoff program, all the code samples and workshop recordings are available in the member portal.
What Our Community Had to Say
“The feedback and collaboration this week were invaluable. Real-time AI models are critical for search-and-rescue missions, and Intel’s technology really elevated our software’s potential. We’re confident it can make a real difference in the field.” – Jack Murray, Co-Founder and CTO, ParaWave.
“This hackathon was incredible. AI-driven optimizations in cancer cell therapy manufacturing can have a massive impact. With Intel’s support, we’ve got the tools to improve production efficiency and ultimately, patient outcomes.” – Geoffrey Stephens, Founder and CEO, AICella,
“It’s amazing how Intel’s tools have helped us scale our AI-powered shopping assistant. We’re excited to offer a smarter, more personalized shopping experience to customers across ecommerce platforms, starting with Shopify.” – Divya Padhyay, Co-Founder and CEO, The Mango Jelly.
Our Next Adventure: Intel® Liftoff Days
We’re already preparing for the next Intel® Liftoff Days. Mark your calendar- Q2 Liftoff Days will bring new tech, new teams, and more chances to build and learn together.
Keep an eye out for signup links in your inbox soon.
About Intel® Liftoff
Intel® Liftoff for startups is open to early-stage AI and machine learning startups. This free virtual program helps you innovate and scale, no matter where you are in your entrepreneurial journey.
Resources used during Liftoff Days
6th Gen Intel® Xeon® Scalable Processor - Latest generation of enterprise server processors
Intel® Tiber™ AI Cloud - Cloud platform for AI development and deployment
Intel® Gaudi® 2 AI accelerator - High-performance AI training processor designed for deep learning workloads
As a Software Tools Ecosystem Specialist at Intel, I’ve had the privilege of working on the dynamic GenAI initiative. My focus is on driving engagement with the software developer audience. I'm a proud team member of the Intel® Liftoff for Startups and the Developer Engagement, Relations, and Studio team.