Here are some recent projects we've built or contributed to:
Persona ©
A mindfulness, self-growth app (original concept)
https://persona-app.co (released this year)
An interactive webapp where users can take a quiz to get their Persona (a detailed self-profile), using our proprietary analysis techniques and algorithm. Once users have been introduced to their Persona, they can:
track their progress on numerous self-growth goals - we built short- and long-term memory into the agent, using LLM summarisation and intent/goal detection
receive instant AI responses to questions, personally tailored to their unique experiences, goals, and challenges - the bot also moderates conversations to ensure responses are relevant and appropriate, through separate agent steps and custom prompt-engineering
get weekly affirmation e-mail with AI-personalised self-growth practices
We are building this application with: full-stack web development, Javascript scripting, advanced OO programming, and an AI / LLM pipeline to provide conversational chatbot functionality.
VEG3
A vegan AI assistant, with immense research library (founded by VegCatalyst, collaboration)
https://veg3.ai (Sign up for a free trial!) - CEO, Sam Tucker
A dedicated vegan AI, offering 24/7 access to unlimited personalised responses and 50,000+ vegan resources (including scientific research / journals, organised and readily searchable). VEG3 provides support for users at every step of their vegan journey:
Navigating social challenges with confidence
Receiving 24/7 personalised assistance - using custom-built conversational agents and pipelines with LLM tools such as GPT-4, Claude, Gemini
Finding vegan-friendly opportunities (events, jobs, and other resources) as well as community
Discovering delicious vegan recipes and meal plans
Accessing a vast library of scientific research documents via RAG (retrieval-augmented generation), providing quality, relevant resources to users - using vector databases, knowledge graphs
American Speech Classifier
A custom-built classification model for academic journals (in collaboration with University of Arizona)
We built this classification model to:
identify recurring identity-based categories in academic journal texts, as well as the weights ("words") most associated with each category
assist in future annotation of additional articles beyond the training set
facilitate users' ability to find articles relevant to certain identity-based themes, to assist with student and institution research into sociolinguistic data
We built this application with Python scripting and training a custom machine-learning model for classification of text documents.
We can build it.
Even if your idea seems different from the few projects we've selected to share above, we would love to discuss it with you.
We are prepared to guide you through the full development cycle or assist you wherever you are in the process:
Ideation - we'll help refine the idea, making sure we understand your vision fully
Design - we'll discuss which tools are the best fit, decide on a structure, and project timeline and costs
Development - we'll get the idea built out and ready for testing
Testing - we'll run a panel of tests to ensure everything functions as expected (can add on extra security review service)
Launching - we'll announce the software as available, finally putting it in your users' hands!
Refining - we'll check in with post-launch performance and user feedback, making sure the completed project continues to align with your original vision
Fill out this intake form ("contact us") and book a free consultation - we're excited to help!