Lead AI Architect. PhD in AI/Machine Learning.

Hoshi Digital Ltd
London
1 day ago
Create job alert

Rome / London / Fully Remote only for exceptional candidates, with periodic travel to team locationcontract

12-month full-time contract, with high-potential for long-term continuation

April / May 2026


About Hoshi Digital


Hoshi Digital is a London-based digital transformation company founded in 2023 by seasoned professionals with deep roots in enterprise technology delivery. We are now firmly focused on AI product development, building proprietary, production-grade systems for enterprise clients across Europe. We work at the intersection of research and delivery. We build things that actually ship.


About This Engagement


You will be joining a confidential, high-priority project involving the design and delivery of a sophisticated AI platform built around computer vision, multimodal reasoning, forensic image analysis, and hybrid edge-cloud inference. The platform will be trained on a proprietary dataset of over 500,000 labelled real-world images and must operate reliably across both online and offline environments.


The project is structured across five interconnected AI modules, each requiring primary models, fallback strategies and EU-hosted alternatives. It integrates with enterprise CRM infrastructure and is designed for full production deployment within 12 months.


A preliminary technology selection has been made as a starting point. It includes SAM 2, YOLO11-seg, LLaVA-Next, DINOv2, LoFTR, Stereo-IMU, Depth Pro, Milvus, and EU-hosted alternatives via Regolo.ai. This selection represents our best current thinking — not a locked specification. The AI landscape is moving fast enough that some of these choices may be superseded by better tools before or during the build. We expect the Lead AI Architect to treat this stack as a hypothesis, validate it rigorously, and replace any component where a superior alternative exists at the time of implementation.


The complexity is real. The timeline is fixed. The dataset is substantial. The stakes are high.


The Role


As Lead AI Architect, you are responsible for the overall technical architecture, model orchestration, and delivery integrity of the platform. You will make the critical decisions — on model selection, edge vs cloud inference, data pipeline design, fallback strategies, and EU compliance — and you will own the consequences of those decisions through to production.


You will lead a hybrid team of five senior engineers and various specialised AI agents that will be created in the initial phase of the project to assist the team.


You will also serve as the primary liaison with an academic research partner responsible for the R&D and certification of one of the platform's core security modules.


Beyond the technical, you are the person who holds the team together. You create the conditions for constructive disagreement, mutual accountability, and honest communication. You are comfortable having difficult conversations and skilled at bringing people to a conclusion without leaving casualties. You are the "enabler", not the "boss".


This is not a role for someone who has only worked in research, or only in engineering. You need to have done both — published work and shipped systems. You understand that in a fast-moving AI landscape, the right architecture today may need to be partially rebuilt in six months, and you plan accordingly.


What You Will Be Responsible For


  • Reviewing, validating and where necessary replacing the preliminary technology selection based on the state of the AI landscape at the time of implementation


  • Designing and owning end-to-end technical architecture across five AI modules and two operational applications


  • Orchestrating the full model stack across vision, multimodal reasoning, forensic analysis, 3D metrology and vector search


  • Leading the data pipeline strategy for a 500,000 image proprietary dataset — anonymisation, auto-labelling, structuring and training readiness


  • Defining and managing fallback model strategies for offline and low-connectivity environments


  • Ensuring EU data residency and compliance requirements are met across all model and infrastructure choices


  • Liaising with an academic research partner to translate R&D output into production-ready components


  • Mentoring and guiding the Computer Vision Engineer, ML Engineer, Senior Mobile Developer, Backend and Data Engineer, and Salesforce Integration Specialist


  • Maintaining architectural coherence as requirements evolve and as the AI landscape shifts during the build period


  • Building and sustaining a team culture of constructive challenge, accountability and mutual respect


What Makes This Role Exceptional


The technical complexity of this project is not manufactured to sound impressive on a job description. You will be architecting a real production system, on a real timeline, with a real proprietary dataset, for a real enterprise client. Every architectural decision you make will have measurable consequences.


You will have genuine autonomy, a capable team, an academic partner for the research-heavy components, and the freedom to challenge every technology choice on the table. If you are the kind of person who finds complexity energising rather than exhausting, this role was built for you.


How We Hire


We do not recruit on CVs alone. Shortlisted candidates will complete a structured technical and behavioural assessment designed to evaluate depth of knowledge, problem-solving approach, and team fit. We are assembling a team that works well together, not just a collection of skilled individuals.


What We Are Looking For


  • 10 to 15 years of hands-on AI and machine learning experience, with demonstrable production deployments at scale


  • PhD in Artificial Intelligence, Computer Science or Machine Learning strongly preferred; publications in computer vision, multimodal AI or related fields are a significant advantage


  • Broad and current knowledge of the AI model landscape — you know what exists, what is emerging, and how to evaluate competing tools objectively


  • Experience designing hybrid edge and cloud inference architectures with offline-first constraints


  • Proven ability to lead senior technical teams through complex, fast-moving projects where the ground shifts beneath you


  • Comfortable operating at the intersection of academic research partnerships and commercial delivery


  • A natural ability to facilitate difficult technical conversations and bring a team to constructive conclusions without ego or politics


  • Fluency in English required; Italian is a strong advantage


  • A genuine belief that AI agents are collaborators, neither a threat to skilled engineers, nor a tool.


Related Jobs

View all jobs

Lead AI Architect. PhD in AI/Machine Learning.

Machine Learning Engineer (Manager)

Principal Data Scientist

AI Data Engineer

AI Data Engineer

Senior Data Engineer

Subscribe to Future Tech Insights for the latest jobs & insights, direct to your inbox.

By subscribing, you agree to our privacy policy and terms of service.

Industry Insights

Discover insightful articles, industry insights, expert tips, and curated resources.

Careers

How Many Machine Learning Tools Do You Need to Know to Get a Machine Learning Job?

Machine learning is one of the most exciting and rapidly growing areas of tech. But for job seekers it can also feel like a maze of tools, frameworks and platforms. One job advert wants TensorFlow and Keras. Another mentions PyTorch, scikit-learn and Spark. A third lists Mlflow, Docker, Kubernetes and more. With so many names out there, it’s easy to fall into the trap of thinking you must learn everything just to be competitive. Here’s the honest truth most machine learning hiring managers won’t say out loud: 👉 They don’t hire you because you know every tool. They hire you because you can solve real problems with the tools you know. Tools are important — no doubt — but context, judgement and outcomes matter far more. So how many machine learning tools do you actually need to know to get a job? For most job seekers, the real number is far smaller than you think — and more logically grouped. This guide breaks down exactly what employers expect, which tools are core, which are role-specific, and how to structure your learning for real career results.

Jobs

What Hiring Managers Look for First in Machine Learning Job Applications (UK Guide)

Whether you’re applying for machine learning engineer, applied scientist, research scientist, ML Ops or data scientist roles, hiring managers scan applications quickly — often making decisions before they’ve read beyond the top third of your CV. In the competitive UK market, it’s not enough to list skills. You must send clear signals of relevance, delivery, impact, reasoning and readiness for production — and do it within the first few lines of your CV or portfolio. This guide walks you through exactly what hiring managers look for first in machine learning applications, how they evaluate CVs and portfolios, and what you can do to improve your chances of getting shortlisted at every stage — from your CV and LinkedIn profile to your cover letter and project portfolio.

Careers Jobs

MLOps Jobs in the UK: The Complete Career Guide for Machine Learning Professionals

Machine learning has moved from experimentation to production at scale. As a result, MLOps jobs have become some of the most in-demand and best-paid roles in the UK tech market. For job seekers with experience in machine learning, data science, software engineering or cloud infrastructure, MLOps represents a powerful career pivot or progression. This guide is designed to help you understand what MLOps roles involve, which skills employers are hiring for, how to transition into MLOps, salary expectations in the UK, and how to land your next role using specialist platforms like MachineLearningJobs.co.uk.

🍪 We use cookies to enhance your browsing experience on our website. By continuing to use this site, you consent to the use of cookies as described in our Cookie Policy. You can review our Cookie Policy for more information.