Tudor Jianu

AI & Autonomous Systems Engineer

About

AI & Autonomous Systems Engineer specialising in scalable backend architecture, MLOps infrastructure, and applied machine learning systems. Experienced in designing secure, production-grade platforms across HealthTech, CivicTech, robotics, and scientific computing domains. Proven track record of translating complex domain challenges into validated technical solutions through cloud-native deployment, model versioning, automation pipelines, and robust data architectures. Strong background in Python, Go, PostgreSQL, and modern web frameworks, with hands-on experience delivering end-to-end systems in multi-stakeholder and publicly funded innovation environments.

Skills

Soft Skills
Commuinication
Artificial Intelligence & Machine Learning
Deep Learning (DL)Machine Learning (ML)Computer VisionNeural Networks (CNNs, RNNs, Transformers)Reinforcement Learning (RL)
Data Science & Analytics
Data VisualizationStatistical AnalysisExploratory Data Analysis (EDA)Data Wrangling & PreprocessingFeature Engineering & SelectionData Pipeline DevelopmentDatabase Systems (SQL & NoSQL)
Software Engineering
Object-Oriented Design (OOD)Model-View-Controller (MVC) ArchitectureDatabase Modelling (ER diagrams, normalization, schemas)Debugging & Unit/Integration TestingContinuous Integration & Continuous Deployment (CI/CD)
Web Development
Front-End Development (HTML, CSS, JavaScript, React)Back-End Development (Node.js, Express, MongoDB)Responsive Web Design & UI/UX PrinciplesDatabase Management & Integration (SQL, NoSQL, ORM)Web Security & Authentication (Encryption, OAuth)
Programming Languages
GoPythonJavaScript/TypeScriptSQLBashHTML/CSS
Tools & Platforms
PyTorchTensorFlowLinux (CLI)Docker (Containerization)Git (Version Control)Jupyter Notebooks (Interactive Environments)Mongoose (MongoDB ORM)Linux (Arch Linux, Debian-based systems)

Experience

Apr 2025 – Present

AI & Autonomous Systems Engineer

Virtual Engineering Centre (VEC)
  • Delivered applied AI, data, and automation solutions across HealthTech, CivicTech, Robotics, and Materials Science programmes within publicly funded innovation initiatives
  • Architected secure, production-ready systems spanning full-stack platforms, MLOps infrastructure, and cloud-native deployments
  • Designed and implemented relational data models, APIs, and role-based access control (RBAC) frameworks to support multi-stakeholder environments
  • Established automated CI/CD pipelines, containerised deployments, and model versioning systems to enable reproducible and scalable software delivery
  • Collaborated directly with entrepreneurs, researchers, public sector partners, and SMEs to translate domain problems into validated technical architectures
  • Mentored engineers and innovation teams in software engineering best practices, system design, and scalable architecture
  • Delivered technical workshops and live demonstrations to SMEs, academic audiences, and public sector stakeholders, including showcasing humanoid robotics capabilities, state-of-the-art computer vision systems, and human interaction models at an Emerging Markets workshop
  • Contributed to regional innovation programmes supporting NHS transformation, social care modernisation, open data infrastructure, and advanced materials research
  • Operated at the intersection of AI engineering, backend systems architecture, and autonomous systems deployment in real-world environments
Oct 2021 – Nov 2025

PhD Researcher in Artificial Intelligence and Endovascular Navigation

University of Liverpool
  • Pioneered AI-driven autonomous endovascular navigation research, focusing on reinforcement learning (RL) and deep learning (DL) methods for guidewire control in complex vascular environments, with a focus on real-world clinical translation.
  • Designed and developed [CathSim](https://airvlab.github.io/cathsim/), an open-source object-oriented Python library for reinforcement learning-based endovascular guidewire navigation. Integrated high-fidelity aortic arch models, continuous guidewire shape representations, and simulated imaging for sim-to-real transfer. Deployed scalable RL training pipelines on HPC environments using Docker containerization.
  • Invented and implemented Splineformer, a novel transformer-based deep learning model using B-spline shape representations to predict guidewire trajectories. Achieved 50% success rate in autonomous cannulation of the Brachiocephalic Artery (BCA).
  • Created and publicly released [Guide3D](https://airvlab.github.io/guide3d/), the first large-scale, annotated dataset of biplanar X-ray sequences for endovascular navigation. Leveraged CVAT for precise annotations and structured data for PyTorch pipelines, enabling research in tool tracking, domain adaptation, and AI-based medical imaging.
  • Advanced autonomous navigation algorithms by developing and optimizing multi-modal reinforcement learning (MMRL) approaches that fuse shape-based and image-based modalities, significantly improving performance in anatomically complex scenarios.
  • Presented original research at international conferences, effectively communicating novel AI methodologies for endovascular robotics to multidisciplinary audiences, including AI researchers, clinicians, and engineers.
  • Mentored MSc and undergraduate students, fostering technical development and cross-disciplinary innovation in AI for healthcare, resulting in increased research output and collaborative publications.
Sept 2024 – Present

Web Developer

University of Liverpool

Contributed to the design and development of a full-stack web application, utilizing modern frameworks and tools to create a functional and user-friendly solution.

  • Designed and optimized backend functionalities utilizing MongoDB and Mongoose ORM, focusing on robust schema design and efficient query optimization to ensure scalable and performant NoSQL data management.
  • Engineered responsive, accessible, and intuitive UI components using React, adhering to modern front-end best practices and state management techniques. Collaboratively enhanced user experiences through consistent interface design and functionality.
  • Leveraged Next.js for Server-Side Rendering (SSR) and Static Site Generation (SSG), significantly improving application performance through features like automatic code splitting and Incremental Static Regeneration (ISR).
  • Implemented comprehensive unit and integration testing using Jest. Integrated continuous integration and continuous deployment (CI/CD) pipelines and debugging tools to ensure high standards of application reliability, maintainability, and quality assurance.
Oct 2021 – Present

Teaching Assistant

University of Liverpool
  • Computer Vision (COMP338): Provided instructional support for coursework in image processing, feature extraction, and machine learning techniques, facilitating student understanding and practical application in computer vision.
  • Robotic Perception and Manipulation (COMP341): Guided students through concepts in robotic control algorithms, perception systems, and manipulation tasks, promoting hands-on experience and effective problem-solving skills in robotics.
Jan 2023 – Apr 2023

Technical Lead

Digital Theme UK-Ukraine Twinning Initiative
  • Designed and implemented technical solutions ensuring a seamless and robust virtual conference experience.
  • Coordinated closely with academic institutions in the UK and Ukraine, fostering international knowledge exchange and technical cooperation.

Projects

Digital Commons Data Portal

Liverpool City Region Civic Data Cooperative
Jan 2025 – Feb 2026

A regional open data platform enabling secure dataset sharing across the Liverpool City Region to strengthen academic collaboration, cross-sector research, and public sector transparency.

  • Contributed to the deployment and customisation of a CKAN-based regional data platform supporting 17 organisations
  • Designed and implemented a production data visualisation plugin, enabling interactive exploration of 45 datasets and 2,737 resources
  • Deployed and configured a production-grade Python environment to support custom extensions and scalable data workflows
  • Supported automated data ingestion and update pipelines, improving data freshness and reducing manual processing
  • Helped operationalise a shared regional data infrastructure used by local authorities and national government bodies
PythonCKANData VisualisationOpen DataServer Deployment

An AI-driven SaaS business intelligence platform providing SMEs with 24/7 actionable insights, performance management and decision support by integrating accounting data from systems like Xero and QuickBooks. The platform helps small business owners and advisors improve cash flow, productivity and profitability. :contentReference[oaicite:1]{index=1}

  • Conducted a security assessment of the existing platform, identified vulnerabilities, and implemented fixes to fortify application and data security
  • Strengthened production reliability and infrastructure through platform deployment and operational hardening
  • Architected and established the foundation for a second major platform version, setting up scalable backend services and frameworks
  • Implemented secure data workflows connecting Xero with the internal PostgreSQL database, enabling synchronized financial data processing
  • Built and integrated authentication and role-based access control (RBAC) using Supabase to support multi-tenant access patterns
  • Developed transactional email workflows within a Next.js frontend, and deployed the stack via Vercel and Supabase
Next.jsPostgreSQLSupabaseXero APIRBACVercelSecurity HardeningSaaS

Shared Lives Plus Digital Platform (CHI-Zone)

CHI-Zone, Civic Health Innovation Labs, University of Liverpool

A community-focused digital platform developed under the CHI-Zone programme, funded by the Liverpool City Region Life Sciences Innovation Zone and anchored by the University of Liverpool’s Civic Health Innovation Labs (CHIL). The system modernises Shared Lives social care workflows, automating carer matching and case management to improve placement efficiency and support person-centred community care.

  • Architected and delivered a full-stack HealthTech platform using Next.js and PostgreSQL to digitise and streamline adult social care workflows
  • Translated person-centred care processes into a scalable relational data model through structured stakeholder discovery and domain analysis
  • Designed a normalised PostgreSQL schema with indexed search optimisation, enforced referential integrity, and controlled schema migrations
  • Integrated secure cloud infrastructure using Amazon RDS and AWS S3 to support compliant storage of care documentation and structured case data
  • Designed and implemented a RESTful API layer enabling structured carer-to-individual matching logic
  • Implemented authentication, role-based access control (RBAC), and custom permission layers aligned with safeguarding and multi-stakeholder governance requirements
  • Established automated unit testing using Vitest with in-memory PGlite database mocking to ensure deterministic and maintainable releases
  • Developed database seeding strategies and environment configuration to support CI/CD and reproducible deployments
  • Contributed to service blueprinting and rapid prototyping to reduce administrative burden, enabling leaner processes and faster access to carers for individuals in need
HealthTechSocial Care TechnologyCommunity Care SystemsNext.jsPostgreSQLDatabase ArchitectureAmazon RDSAWS S3REST API DesignRBACCloud InfrastructureCI/CDTest-Driven DevelopmentStakeholder EngagementService Design

Citizen First – MySleepHealth (CHI-Zone)

CHI-Zone, Civic Health Innovation Labs, University of Liverpool

A venture supported under the Citizen First programme, empowering lived-experience entrepreneurs in the Liverpool City Region to build health-focused businesses. MySleepHealth addresses sleep-related health challenges through a digital platform designed to improve user wellbeing and access to structured support.

  • Designed and implemented the backend infrastructure for a Flutter and Supabase-based HealthTech application
  • Worked directly with the founder through iterative requirement discovery to translate lived-experience insights into scalable technical architecture
  • Developed secure data models and backend services within Supabase, enabling structured user data handling and application logic
  • Implemented authentication and access control to support secure, multi-user health data interactions
  • Supported rapid MVP development to accelerate validation, enabling the entrepreneur to focus on product-market fit and community impact
  • Contributed technical expertise within a publicly funded innovation programme aimed at empowering hidden entrepreneurs to solve real health and care challenges
HealthTechStartup InfrastructureSupabaseFlutter BackendBackend ArchitectureMVP DevelopmentFounder CollaborationProduct ValidationCommunity InnovationSocial Impact Technology

Citizen First – Braywell Robotics (CHI-Zone)

CHI-Zone, Civic Health Innovation Labs, University of Liverpool

A HealthTech venture supported under the Citizen First programme focused on improving infection control and sanitation validation in NHS hospital environments through robotics and automation. The solution combines a mobile robotic base with a continuum manipulator to inspect and validate hygiene compliance in complex clinical settings.

  • Provided technical advisory support to validate the feasibility of a robotic sanitation verification system for NHS hospital environments
  • Supported early-stage technical architecture and systems thinking for integrating a continuum robot with a mobile robotic platform
  • Advised on validation workflows aligned with hospital infection prevention and control standards
  • Contributed to technical risk assessment and feasibility evaluation to support product validation and funding readiness
  • Worked closely with the entrepreneur to translate a clinical problem into a scalable HealthTech robotics solution
  • Supported a community-driven innovation initiative aimed at improving operational efficiency and safety within NHS settings
Healthcare RoboticsContinuum RoboticsMedical Device InnovationNHS InnovationHealthTech ValidationTechnical FeasibilitySystems ArchitectureInfection Control TechnologyRobotics in HealthcareInnovation Advisory

LMDS – Automated Machine Learning Deployment Platform

Liverpool Materials Discovery Studio (LMDS)

An internal MLOps platform developed within the Liverpool Materials Discovery Studio (LMDS) to automate versioned deployment of machine learning models for functional materials research. The system enables non-technical users to execute validated models through a secure interface, bridging advanced computational research with accessible tooling.

  • Architected and implemented an end-to-end automated model deployment system enabling version-controlled ML model releases via GitHub tag workflows
  • Built a Go-based API layer orchestrating deployment pipelines, container lifecycle management, and runtime monitoring
  • Configured action runners to automate packaging of Python (scikit-learn) models into Docker images and publish them to a locally hosted container registry
  • Implemented structured model versioning and artifact registration within PostgreSQL to support reproducibility, validation, and parallel model coexistence
  • Designed a repository template using uv to standardise model contracts, enforce parameter schemas, and enable frontend auto-generation of input forms
  • Established secure authentication and role-based access control (RBAC) to manage model publishing, validation, and execution permissions
  • Implemented TCP log streaming from running Docker containers via the Go API to provide real-time execution feedback to the frontend
  • Delivered infrastructure enabling non-technical researchers to run complex ML workflows without direct coding interaction
MLOpsModel VersioningAutomated DeploymentGo APIDockerGitHub ActionsPostgreSQLRBACMachine Learning InfrastructureScikit-learnContainer OrchestrationResearch Software Engineering

VEC Digital Registry System

Virtual Engineering Centre (VEC)

An internal access management platform developed to replace unreliable paper-based building entry and exit logs. The system improves auditability, security, and operational efficiency by digitising visitor registration, staff access, and group bookings in a continuously running production environment.

  • Replaced manual paper logbooks that were prone to incomplete records, poor traceability, and compliance risks with a centralised digital registry system
  • Designed and implemented a cross-platform desktop application using Electron and Svelte with PostgreSQL-backed persistent logging
  • Integrated RFID-based authentication to eliminate manual staff sign-in, reducing entry friction while improving access control accuracy
  • Implemented group booking and bulk check-in/out workflows to address bottlenecks during events and training sessions
  • Enabled real-time occupancy visibility and structured access logs to support safety, compliance, and incident traceability
  • Deployed and maintained the system on a Debian production environment, running 24/7 for nearly a year with minimal downtime
Access Control SystemsOperational DigitisationElectronSveltePostgreSQLRFID IntegrationLinux DeploymentInfrastructure AutomationCompliance TrackingReal-Time Logging Systems

VEC Holiday & Leave Management System

Virtual Engineering Centre (VEC)

An internal leave management platform developed to replace unreliable and manually calculated holiday tracking processes. The system ensures accurate entitlement calculation, structured approval workflows, and full auditability, reducing administrative risk and payroll discrepancies.

  • Replaced inconsistent manual leave tracking with a centralised, rules-driven holiday management system
  • Designed and implemented the full-stack application using SvelteKit and PostgreSQL with a robust relational schema modelling annual leave, sick leave, public holidays, and institution-specific custom holidays
  • Engineered accrual and pro-rata calculation logic to support different contract types, entitlement rules, and carry-over policies
  • Implemented approval workflows with designated approvers and structured decision tracking to ensure governance and transparency
  • Developed integrated team-based holiday calendars to provide organisational visibility and reduce scheduling conflicts
  • Built custom authentication with invitation-based onboarding and two-factor authentication (2FA) to secure sensitive HR data
  • Established automated balance recalculation and audit logging to ensure reliable, traceable leave history
  • Delivered a reliable operational system improving compliance, transparency, and administrative efficiency across the organisation
SvelteKitPostgreSQLWorkflow AutomationRBACTwo-Factor AuthenticationHR Systems EngineeringAccess ControlRelational Database DesignBusiness Logic ModellingOperational Digitisation

Education

Oct 2021 – Present

PhD in Artificial Intelligence within Endovascular Navigation

University of Liverpool
Sept 2020 – Sept 2021
Distinction

Masters of Science in Computer Science

University of Liverpool
Sept 2017 – Aug 2021
First Class Honours

Bachelours of Arts in Business Administration

Coventry University

Volunteering

Oct 2021 – Present

Outreach Team Member

University of Liverpool
Sept 2020 – Sept 2021

President of Data Science and Artificial Intelligence (DSAI) Society

University of Liverpool
  • Established the society from the ground up, setting strategic goals to foster student engagement in Data Science and AI.
  • Organized networking events and hands-on workshops, connecting students with industry professionals and providing practical training.

Interests

Technology
Linux (Arch Linux, Debian-based systems)Self-hosting solutions (Nextcloud)Tiling window managers (qtile)Terminal-based tools (e.g., neovim, tmux, zsh/bash)Scripting and automation (Bash, Python)Open-source contributions and toolsQMK (split keyboard customization)
Sports & Fitness
BoulderingGymnastics

© 2026 Tudor Jianu