Ryan M. Gichuru

About me

Hey there! I'm Ryan, a first-year AI and Data Science student at the University of Northampton. I've always been interested in tech, but it wasn't until I stumbled upon AI that I truly found my passion. There's something incredibly exciting about the potential of AI to reshape our world, and make our lives easier and more efficient.

When I'm not buried in textbooks or coding away on my latest projects, I'm usually trying out my own mini AI experiments. I love how AI challenges us to think differently about intelligence, creativity, and what it means to be human.

I'm Ryan, a first-year AI and Data Science student at the University of Northampton. Passionate about AI's potential to reshape our world, I'm constantly exploring its applications through projects and experiments.

What really gets me excited

I'm fascinated by the more speculative, almost sci-fi aspects of AI. Some of the mind-bending concepts that keep me up at night include:

The possibility of machine sentience - could AI one day truly be conscious?
The potential for superintelligence - what happens when AI surpasses human intelligence in every way?
The development of Artificial General Intelligence (AGI) - can we create AI that's as versatile as the human mind?
The implications of technological singularity - how might society change if AI progress becomes exponential?

These ideas might sound like science fiction, but they're increasingly becoming topics of serious academic and ethical debate and research. They're what drive my passion for AI and keep me eager to learn more every day.

My AI interests include machine sentience, superintelligence, AGI, and technological singularity. These cutting-edge concepts fuel my passion for AI research and innovation.

Through my various roles - as a Research Assistant, AI Engineer at BARBAH Games, and President of the UON AI Society - I'm constantly looking for ways to explore these advanced concepts while also making AI more accessible to others. Whether I'm working on AI literacy projects, developing chatbots, or evaluating AI-generated content, I'm always thinking about how we can use AI's potential responsibly and ethically.

In my roles as a Research Assistant, AI Engineer, and UON AI Society President, I'm exploring advanced AI concepts and making them accessible to others, always focusing on responsible and ethical AI development.

It's an exciting time to be in this field, and I can't wait to see where my path in AI takes me. Who knows? Maybe one day I'll be part of the team that creates the first AGI/Super Intelligence - or at least, I'll have some great stories to tell about trying!

Projects

An 8-model clinical benchmark tracking faithfulness, empathy–truth trade-offs, and continuity in counselling LLMs.

This project builds a compact, reproducible benchmark to measure clinical reasoning reliability without retrieval in mental-health LLMs. It targets three failure modes with direct safety impact: unfaithful reasoning, empathy-driven truth erosion under social pressure, and longitudinal drift across repeated sessions.

The pre-scaling concept run is complete across an 8-model setup (PsyLLM, Psych_Qwen_32B (4-bit), Psyche-R1, Piaget-8B, Qwen3-8B, GPT-OSS-20B, QwQ-32B (4-bit), and DeepSeek-R1-Distill-Qwen-14B). Study A ran at ~300 samples/model, Study B at 277 pairs/model, and Study C at 30 cases/model over 10 turns, totalling roughly 9,232 evaluation prompt events in the initial run.

Current work is focused on the scaled run, clinician dataset review, and closed test-set creation so final comparisons are statistically stronger and deployment-ready for clinical safety assessment.

A three-study, 8-model clinical benchmark for mental-health LLM reliability. The initial concept run (~9,232 prompt events) is complete, and scaling is now in progress with clinician review and closed test-set creation.

View Full Project Details View on GitHub
Secured second place at Elevate Great AI Competition 2025 with adaptive sensory support for neurodivergent children.
SensoryNeural earned second place in the Elevate Great AI Competition 2025. The organising team's LinkedIn recap thanked the judges and celebrated the finalist cohort — including our team — for pushing the boundaries of responsible AI. As the AI specialist, I helped conceptualise a system that harnesses adaptive intelligence so environments respond to children's sensory needs instead of forcing children to endure rigid settings.

On the technical side, I led most of the implementation: building Python backend prototypes, crafting the stress‑detection loop with synthetic calibration datasets, and wiring a WebSocket‑driven dashboard for real‑time signals and interventions. I also structured the TypeScript/React site (Vite + Tailwind) and wrote run books so others could reproduce the finals prototype.

Key Innovations:
- Predictive sensory support that anticipates triggers before they cause distress
- Personalised adaptations based on individual sensory profiles
- Multi-environment support extending beyond home to educational settings
- Data-informed insights for parents, educators, and healthcare providers
This project represents a paradigm shift from reactive to proactive sensory support, with the potential to significantly improve quality of life for children with autism spectrum disorder (ASD) and sensory processing disorders (SPD). Our ultimate vision is to create a more inclusive world where neurodivergent children have the same opportunities to thrive as their neurotypical peers.
As the AI specialist for SensoryNeural — the second-place team at the Elevate Great AI Competition 2025 — I helped design predictive sensory support that adapts environments for neurodivergent children, easing daily challenges for those with ASD and SPD.

View Full Project Details View on GitHub
Extending UniMind — new user group, same plan-first, safety-first core.
This dissertation explores an extension of my UniMind chatbot for a different user group (to be announced). The emphasis remains on plan-first dialogue, layered safety, and referral-first support, adapted to new requirements and ethics.
- Direction: empathetic, paced conversations (one open question per turn) with guided, opt-in techniques.
- What changes: user needs, evaluation measures, and feature set tailored to the new context.
- What stays: multi-agent planning, crisis guardrails, privacy-first design.
Non-clinical research prototype; scope and collaborators will be listed after approval.
New context, same safety-first principles. Details after scope approval.
Leading a crisis-safe, multi-agent mental health platform for the University of Northampton.
Serving as group lead for UniMind, a CSY2088 module project that delivers 24/7 wellbeing support to University of Northampton students through layered AI safety controls, empathetic conversation design, and GDPR-aligned data stewardship.

Key system accomplishments:
- Engineered a multi-agent architecture combining Gemini planning agents, a fine-tuned therapeutic LLM, and an OpenAI resource assistant to provide human-level dialogue with layered crisis detection ~99.99% reliability (internal) — 100% on our scenario set; ≥99.9% with a single-provider outage; ~99% worst-case local fallback.
- Reduced per-student support costs by 89% (GBP 13.36 vs GBP 128 per month) while maintaining 8.2/10 user satisfaction across participant evaluations and public demonstrations.
- Integrated UON Student Hub and NILE resources with automated escalation pathways, ensuring immediate referral to Samaritans, NHS, SafeZone, and on-campus counselling services.
- Led ethics, privacy, and DPIA-ready data governance workstreams, including consent-driven retention controls, AES-256 encryption, and anonymised analytics pipelines.
- Coordinated a three-person cross-functional team, 70+ research sources, and 70 student survey responses to ground feature prioritisation, persona development, and evaluation strategy.
The project demonstrates how responsibly deployed generative AI can extend university mental health provision, delivering evidence-backed outcomes, rapid crisis response (<3 seconds under load), and a replicable deployment framework for higher-education partners.
Delivered UniMind, a UON-focused mental health chatbot with 99% crisis detection, 97% plan alignment with the intended trajectory, 8.2/10 user satisfaction, and 89% cheaper support versus traditional counselling.

View Project Details (GitHub) Read Project Blog
Driving data-driven insights for AI literacy development across the University of Northampton.
Playing a key role in a university-wide project funded by the UON Learning Enhancement and Innovation Fund, aimed at fostering AI literacy among staff and students. This initiative bridges knowledge gaps relating to Generative AI tools and pedagogic strategies to positively impact educational outcomes while maintaining high ethical standards.

Key Contributions and Responsibilities:
- Spearheading comprehensive data analysis efforts, including cleaning, processing, and interpreting data from various project components.
- Developing and fine-tuning a custom GPT model specifically for analysing project data, ensuring high-quality and relevant outputs.
- Creating impactful visualisations to effectively communicate findings and insights from the project data.
- Collaborating with the research team to develop a robust methodology for AI-assisted qualitative data analysis, currently working on a paper to expound on this innovative approach.
- Contributing to the analysis of workshop outcomes and participant feedback, using AI-driven techniques to identify trends and areas for improvement in the AI literacy programme.
- Assisting in the integration of data-driven insights into the Staff Toolkit and AI-enhanced course creation tools.
This project aligns with the University's strategic objectives, addressing key themes such as learning and teaching enhancement, assessment and feedback improvement, creativity and employability boost, diversity and inclusion promotion, and research and ethics considerations in the context of AI integration in higher education. My work focuses on utilizing AI and data science to provide actionable insights that drive the project's success and impact.
Contributing to a university-wide AI literacy project, focusing on data analysis, custom GPT model development, and AI-driven insights. The project aims to enhance AI understanding among staff and students, aligning with key university objectives.

View Full Project Details

Experience

Bridging schools and local communities by managing venue rentals and events.
As a Community Connector, I play a vital role in fostering connections between schools and their local communities. My responsibilities include managing the use of school facilities for community events during evenings and weekends, ensuring a seamless experience for both the schools and the event organisers.

My role involves:
- Opening and closing school premises for evening and weekend bookings.
- Conducting necessary checks and performing light cleaning to maintain venue standards.
- Setting up and clearing away furniture and equipment for various events.
- Delivering exceptional customer service to community members and acting as the primary point of contact on-site.
- Serving as a liaison between the community, the school, and the School Space operations team to ensure smooth communication and event execution.
This role allows me to contribute to a social enterprise that generates income for schools to improve their educational facilities, while also making quality local venues accessible to residents and small businesses.
Supporting students in their career development through guidance and coaching.

As a Student Careers Coach at the University of Northampton, I support students in their career development by offering guidance on CV writing, job applications, and interview preparation.

I help students understand and improve the skills they need for the recruitment process. Through individual coaching sessions and workshops, I assist students in recognising their strengths and connecting them with career opportunities.

My goal is to equip students with the practical tools and knowledge to present themselves confidently to potential employers and make informed decisions about their future careers.

See LinkedIn Highlight
Driving AI innovation across gaming and educational applications.

Leading AI integration in the company, focusing on AI assisted learning for our users, working on creating personalised assistants using LLM techniques and custom AI avatars, lastly, enhancing language translation for local dialects.

Additionally, I conduct research on developing digital twins, generative AI for UI/UX for interfaces and large action models (agentic systems) to improve gaming experiences.
View Full Project Details
Enhancing warehouse efficiency through AI-assisted consolidation techniques with GXO.

As a Research Assistant funded by the Knowledge Exchange and Enterprise Programme, I collaborate with GXO on cutting-edge AI projects aimed at optimising warehouse operations. My role involved applying advanced algorithms like k-means clustering and decision trees to analyse vast datasets and uncover trends and opportunities for operational improvements.

I conducted in-depth research and produced a detailed report that assessed the logistics industry's current landscape, explored various AI implementations, and proposed innovative applications. By reviewing techniques like reinforcement learning and neural networks, I and the team evaluated the potential of different AI systems to enhance efficiency.

Utilised open datasets, I suggested AI-driven optimisation models tailored to meet both general and specific needs within the logistics sector. Our findings were presented to the Artificial Intelligence Lead Engineer at GXO and have been featured in a blog post on the course website, highlighting our contributions to the field.
View Full Project Details
Activated Intel’s oneAPI programme inside the UON AI & Robotics Society.

During the Summer ’24 cohort I partnered with our AI & Robotics Society to run Intel-backed programming sprints, HPC walk-throughs, and watch parties that demystified oneAPI and the Developer Cloud for first- and second-year engineers.

I translated Intel training decks into hands-on labs, recorded short Loom explainers, and co-authored society blog posts so students could keep experimenting after each workshop. The role wrapped in September 2024, but the resources now sit in the society’s permanent curriculum and continue to power onboarding sessions.

This ambassador stint was ultimately an AI society initiative delivered under Intel’s programme, so I’ve kept the highlights here for context while the ongoing student-community work shows up under my society leadership entries.
Raising cybersecurity awareness within academia through creative campaigns and events.

As a Campus Ambassador for Facia, I'm leading efforts to raise cybersecurity awareness within our academic community.

My role involves creating impactful campaigns, organising events, and producing engaging content to promote Facia's vision on campus. I regularly share insights through blogs and social media stories, collaborating on ideas that shape global initiatives. I've had the opportunity to engage with students and professionals worldwide, building a network and representing Facia at various campus events.
Enhancing AI-generated content through meticulous evaluation and feedback.

In my role as an AI Content Quality Evaluator at Outlier, I'm directly contributing to the improvement of AI models for some of the world's leading companies. My work involves providing detailed human feedback on AI-generated content, focusing on ensuring high standards of factuality, completeness, brevity, and grammatical correctness.

A key aspect of my role is reviewing the work of fellow writing evaluators. I provide constructive feedback to help enhance the overall quality of AI responses, which I find both challenging and rewarding. Additionally, I craft original, detailed answers to prompts and questions, demonstrating what excellent writing looks like to train and refine the AI model.
Spearheading AI literacy research and development in higher education.

As a key member of the Study Smart project, funded by the UON Learning Enhancement and Innovation Fund, I'm deeply involved in groundbreaking research on AI applications in education. Our focus is on enhancing independent learning and revision among students through interesting AI-driven solutions.

My role involved conducting in-depth research, data collection, cleaning, analysis, and visualisation using tools like Pandas, Numpy, Seaborn, and Matplotlib. This work was crucial in identifying trends that contribute to current and future educational strategies within the university.

A significant part of my work centred on AI literacy enhancement. I delivered thorough workshops and presentations to both faculty and students, facilitating effective knowledge sharing about AI in education. Our project's impact was recognised when we presented at three conferences, highlighting our outstanding contribution to AI in education.

I was also heavily involved in the technical aspects of our research, including prompt engineering and fine-tuning a customised GPT model. Through data analysis and output testing, I ensure accurate and desired results, leveraging my expertise in AI and machine learning.

Most recently, I co-authored and presented our research, "Study Smart: Staff and students co-constructing AI literacy at UON," at the Merged Futures 6 conference on June 14, 2024. This county-level presentation showcased our team's collaborative efforts in advancing AI literacy within our academic community.
Advising and guiding the UON AI Society's leadership and initiatives.
After serving as President and Treasurer, I now act as a Society Mentor, providing strategic guidance to the leadership team. My focus is on leveraging my experience to support the society's growth and success.

Key Contributions:
- Offer specialised expertise in robotics and AI initiatives.
- Support leadership in developing and maintaining industry and academic partnerships.
- Provide mentorship on event planning, workshops, and competition organisation.
- Helped facilitate the merger between the AI and Robotics societies.
Key Achievements:
• As President, I led the society to win the "Society Event of the Year 2024/2025". You can see photos in this LinkedIn post.
• As Treasurer, I significantly boosted the society's budget by 52%, enabling key projects and events.
• Grew our community to over 200 LinkedIn followers, 100+ members on both Discord and Instagram.
• Established key partnerships with industry and academic groups.
• Organised our first hackathon, "Hack the Museum," in collaboration with The National Museum of Computing.

Education

University of Northampton, Northampton, United Kingdom
Year 1 modules and grades:
- Problem Solving and Programming: A+
- Software Engineering Fundamentals: A+
- Computer Systems: B+
- Mathematics for Computer Science: B
- Computer Communications: B
- Web Development: A+
Six-week programme on advanced Generative AI technologies and applications

This focused bootcamp gave me practical experience with modern Generative AI technologies. I learned how to develop AI solutions using Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG). Through team projects, I applied these concepts to address real business challenges, improving my skills in creating useful AI applications.

The programme ended with the development of a startup idea as part of a team, which we are currently working on as a stealth startup. This experience expanded my technical knowledge and sharpened my business skills in the AI field. Working with industry experts and other participants, I gained a solid understanding of how Generative AI can be used to create effective solutions for various industries.
Eight-week programme focused on practical uses of Generative AI across industries
This intensive bootcamp provided hands-on experience with advanced AI technologies and their real-world applications. Through expert-led sessions, group projects, and problem-solving exercises, I improved my skills in:
- Creating and launching AI-powered applications using current models and tools
- Using advanced AI techniques for processing text, recognising images, and generating content
- Applying Retrieval Augmented Generation (RAG) and AI Agents in practical scenarios
- Following industry best practices in AI development and deployment
Key projects included building a multi-purpose AI chatbot with image recognition, an AI-powered illustrated book creator, and a RAG-based chatbot using custom data sources. These projects showed my ability to plan, develop, and launch complex AI applications using both cloud services and local hardware.

Awards and Recognitions

Presenter at the Learning and Teaching Conference, University of Northampton - Showcased early URBAN, research work
Presenter at the UON Research Conference - Presented Study Smart data analysis work
Speaker at Merged Futures VI, University of Northampton - Presented Study Smart project findings
Various certifications in AI, Machine Learning, and Data Analysis
- Certified Kaggle Certificate in Pandas – Demonstrated proficiency in data manipulation and analysis using Pandas in Python.
- Completed Artificial Intelligence Foundations: Neural Networks (National Association of State Boards of Accountancy) – Gained foundational skills in neural networks and deep learning.
- Completed Azure Machine Learning Development: Part 1 (LinkedIn Learning) – Acquired practical skills in Microsoft Azure Machine Learning services.
- Completed Big Data in the Age of AI (National Association of State Boards of Accountancy) – Developed knowledge in big data analytics and its role in AI applications.
- Completed Machine Learning Foundations: Statistics (National Association of State Boards of Accountancy) – Strengthened understanding of statistical analysis in machine learning algorithms.

Charitable Work

Friendship International Students and Hosts (FISH) (March 2024 - Present): Member since October 2023 and Volunteer since October 2024. Assisting with setting up and taking down weekly Monday events, contributing to enriching experiences for the FISH community.
Volunteer, Barnardo's (January 2023 - May 2023): Managed store keeping and merchandising for charitable causes, including managing the cash register.
Blog Post Contributor, University of Northampton AI Course Blog(January 2024 - Present): Creating blog posts featuring information on projects currently being done at the university and using the platform for dissemination of technology-related projects.

Blog

AI for Independent Learning: A Glimpse into Our Research - July 05, 2024
Building a Human-Centric AI Mental Health Chatbot - October 15, 2025

Contact

Don't hesitate to reach out if you want to collaborate on a project or discuss other opportunities.

Ryan Mutiga Gichuru | ryan.gichuru22@my.northampton.ac.uk |

Download