AI Systems Architect

About UltaHost
UltaHost is a fast-growing global web hosting and cloud infrastructure company delivering high-performance, reliable, and scalable technology solutions to customers worldwide. Since our founding in 2018, we have grown rapidly across global markets, supporting entrepreneurs, developers, agencies, startups, SaaS companies, and businesses with modern hosting and cloud infrastructure.

Our services span VPS hosting, dedicated servers, shared hosting, domains, game servers, and cloud-based infrastructure across a growing international footprint.

As UltaHost continues to scale, Artificial Intelligence is becoming an increasingly important part of our infrastructure, products, customer experience, and internal operations.

We are investing in: GPU-powered infrastructure, private and self-hosted LLM environments, commercial and open-source AI models, AI-powered applications, intelligent infrastructure automation, RAG and agentic systems, customer-facing AI services, and reusable AI platform capabilities.

We are now looking for an AI Systems Architect to help design and build the technical foundation of this next stage.

Job Overview
We are looking for an AI Systems Architect to help the company adopt AI across departments, reduce manual work, improve sales performance, and build internal AI-powered tools, scripts, platforms, and agents. This is not a traditional AI engineering role. We are looking for a visionary technology leader who can identify opportunities where Artificial Intelligence can fundamentally improve how the company operates, sells, supports customers, develops products, and scales globally.

As AI Systems Architect, you will define and execute the organization’s AI roadmap, lead AI innovation initiatives, build internal AI platforms and intelligent automation systems, and establish our position as an AI-first hosting, cloud, and SaaS company. This person should be both technical and product-minded: able to understand business problems, design practical AI solutions, build prototypes, work with tools such as Lovable and similar AI development platforms, and help promote the company as an AI-first technology brand.

The role is ideal for someone who has strong technical knowledge, good taste in design and user experience, and a clear vision for how AI can improve operations, customer support, sales, marketing, development, and management workflows. The AI Systems Architect will design, build, integrate, deploy, and improve AI systems across UltaHost’s infrastructure and application ecosystem.

This is a highly hands-on technical role operating at the intersection of: GPU infrastructure, Linux systems, Proxmox and KVM virtualization, containers, LLM deployment and inference, AI applications, RAG and vector search, agentic systems, cloud and hosting platforms, backend engineering, infrastructure automation, and production reliability.

You will work directly with our CTO and collaborate closely with Product, Engineering, Infrastructure, Operations, Support, and other business departments. This position does not currently include direct team-management responsibilities. It is an individual-contributor role with significant technical ownership and architectural influence.

We are looking for someone who understands the full production stack:

GPU infrastructure → virtualization and containers → model serving → AI platform services → APIs and integrations → AI applications → monitoring, security, and optimization.

Main Goal of the Role
Your main goal will be to help UltaHost build a practical, scalable, and production-ready AI ecosystem.
You Will

• Design and build infrastructure for running AI and LLM workloads on GPU-enabled environments.
• Deploy, benchmark, optimize, and operate self-hosted and third-party LLMs.
• Build production AI applications, agents, copilots, APIs, RAG systems, and automation.
• Connect AI systems with UltaHost infrastructure, products, customer portals, support systems, databases, and internal tools.
• Explore how technologies such as GPUs, Proxmox, containers, open-source LLMs, vector databases, and modern AI frameworks can become part of UltaHost's technology stack.
• Identify internal processes where AI and deterministic automation can reduce repetitive work and improve efficiency.
• Build reusable AI infrastructure and services that can support multiple future applications instead of isolated one-off experiments.
• Help UltaHost evolve toward an AI-enabled hosting and cloud platform.

The objective is not experimentation for its own sake.
We want working systems that create measurable technical or business value.
Key Responsibilities
GPU & AI Infrastructure

• Design, deploy, configure, and operate infrastructure for LLM and AI workloads using GPU-enabled servers.
• Work with GPU environments and understand practical considerations related to:GPU compute, VRAM capacity, model size, precision, quantization, concurrency, batching, utilization, throughput, latency, thermal and power constraints,and workload allocation.
• Evaluate the hardware and infrastructure requirements of different AI models and use cases.
• Design virtualization strategies for AI workloads using Proxmox VE, KVM, virtual machines, Linux containers, and Docker.
• Configure or support GPU and PCIe passthrough in virtualized environments where appropriate.
• Design secure resource-isolation models for internal and potentially customer-facing AI workloads.
• Contribute to GPU infrastructure capacity planning, availability, monitoring, backup, and disaster-recovery strategies.
• Develop repeatable deployment processes rather than relying on manual server configuration.
• Work with Infrastructure and Engineering teams to create stable environments for development, testing, staging, and production AI workloads.

Proxmox, Virtualization & Platform Engineering

• Design and maintain Proxmox-based infrastructure for AI and general-purpose workloads.
• Work with Proxmox clusters, virtual machines, LXC containers, networking, storage, resource allocation, templates, and backups.
• Design infrastructure with appropriate isolation, high availability, performance, and operational simplicity.
• Automate VM, container, and infrastructure provisioning using APIs, infrastructure-as-code, scripts, or orchestration tools.
• Help standardize deployment patterns across GPU servers and AI application environments.
• Troubleshoot performance, networking, storage, virtualization, and hardware-resource issues.
• Evaluate when workloads should run on bare metal, virtual machines, containers, or orchestration platforms.

LLM Deployment & Inference

• Deploy and operate open-source LLMs on private/self-hosted infrastructure.
• Evaluate and work with inference technologies such as vLLM, Ollama, LiteLLM, Hugging Face, TGI, or comparable platforms.
• Benchmark models based on quality, latency, throughput, VRAM consumption, concurrency, and operating cost.
• Understand and apply inference optimization techniques such as quantization, batching, caching, context management, and model routing where appropriate.
• Compare self-hosted models with commercial APIs such as OpenAI, Anthropic Claude, and Google Gemini and select the appropriate architecture for each use case.
• Design model gateways and reusable inference APIs that can serve multiple UltaHost applications.
• Build resilient model integrations with appropriate fallbacks, retries, rate limits, timeouts, and error handling.

AI Applications, RAG & Agentic Systems

• Design and build production-ready AI applications rather than demonstration-only chat interfaces.
• Develop internal AI assistants connected to authorized company knowledge, documentation, support content, operational runbooks, product information, and other approved data sources.
• Build RAG systems using embeddings, vector databases, metadata filtering, reranking, access controls, and retrieval evaluation.
• Develop AI agents capable of using controlled tools, APIs, databases, and internal services.
• Build multi-step AI workflows and agent orchestration where the complexity is justified by the use case.
• Implement structured outputs, schema validation, tool calling, memory, state management, context management, and execution controls.
• Create internal AI applications for technical and non-technical teams.
• Build customer-facing AI applications or AI-enabled hosting products where strategically relevant.
• Distinguish between tasks that require probabilistic model reasoning and tasks that should be implemented through deterministic software or traditional automation.
• Avoid unrestricted LLM-generated infrastructure actions and use controlled, validated tool interfaces instead.
• Design appropriate human-approval mechanisms for sensitive, unusual, or high-risk AI actions.

Backend Development & Systems Integration

• Build reusable AI services, APIs, and platform components that can support multiple products and departments.
• Integrate AI systems with existing UltaHost platforms, portals, databases, billing systems, support systems, monitoring platforms, and internal applications.
• Work with REST APIs, webhooks, databases, queues, caches, scheduled jobs, and event-driven workflows.
• Work with relational and vector databases such as PostgreSQL/pgvector, Qdrant, or comparable technologies.
• Build backend services using Python, TypeScript/Node.js, or other appropriate technologies.
• Work with Engineering to integrate AI functionality into existing and future customer-facing products.
• Work with Engineering teams to integrate AI functionality into existing and future UltaHost products.
• Ensure that prototypes intended for production are converted into properly tested, documented, and maintainable software.

Infrastructure & Workflow Automation

• Automate infrastructure provisioning, configuration, deployment, and maintenance.
• Work with tools such as:Terraform, Ansible, GitLab CI/CD, GitHub Actions, n8n, scripts, internal APIs, or comparable automation technologies.
• Create idempotent and repeatable infrastructure workflows where possible.
• Identify repetitive operational tasks that can be eliminated or reduced through deterministic automation or AI-assisted workflows.
• Build automations that connect infrastructure, engineering, support, and business systems.
• Evaluate whether a business problem requires AI, standard software development, workflow automation, or a hybrid solution.
• Reduce unnecessary manual dependencies without introducing unsafe or difficult-to-maintain automation.
• Monitor AI infrastructure and applications for performance degradation, model failures, infrastructure issues, and abnormal behavior.

Security, Isolation & AI Governance

• Apply least-privilege principles across infrastructure, applications, services, and agent tools.
• Design secure handling for credentials, secrets, API keys, service accounts, SSH access, and model-provider access.
• Implement authorization boundaries for AI applications and infrastructure actions.
• Consider tenant isolation when AI systems interact with customer data or customer infrastructure.
• Ensure that one customer, user, application, or workload cannot access another customer’s data or resources.
• Design controlled execution mechanisms for agents interacting with servers, APIs, or sensitive systems.
• Implement structured and restricted tool interfaces rather than passing unrestricted model-generated shell commands directly to production environments.
• Establish appropriate:approval gates; action-risk levels, allowlists, deny lists, validation rules, audit trails, and rollback controls.
• Address LLM-specific risks including:prompt injection, indirect prompt injection, hallucination, data leakage, unauthorized tool use, unsafe action generation, and excessive permissions.
• Ensure that AI-generated recommendations and actions are traceable and auditable where required.

Business Automation & AI Adoption

• Work with departments such as Support, Sales, Marketing, Finance, HR, Operations, and Engineering to identify repetitive or inefficient workflows.
• Determine whether each problem is best solved through AI, traditional automation, software development, or a combination.
• Build AI agents, scripts, workflows, dashboards, and internal tools that reduce manual work.
• Integrate AI capabilities into existing business workflows rather than creating disconnected AI demos.
• Evaluate new AI technologies, APIs, open-source projects, models, and platforms and recommend technologies worth adopting.
• Document solutions and help teams understand how to use AI systems effectively.

Required Skills & Experience
We are looking for a strong hands-on technical profile. You do not need to be an expert in every technology listed below, but you should have enough breadth to understand how the components of a production AI system fit together.

Linux, Virtualization & Infrastructure

• Strong Linux administration and troubleshooting skills, preferably with Ubuntu, Debian, or comparable distributions.
• Hands-on experience with KVM-based virtualization or equivalent enterprise virtualization technologies.
• Practical experience with Proxmox VE or closely related virtualization platforms.
• Understanding of virtual machines, Linux containers, virtual networking, storage, resource isolation, templates, snapshots, and backups.
• Strong Docker and containerization knowledge.
• Understanding of networking fundamentals including:TCP/IP, DNS, ports, routing, firewalls, reverse proxies, TLS, and service connectivity.
• Experience operating, debugging, or supporting production infrastructure.

GPU & AI Infrastructure

• Practical understanding of GPU-based AI workloads.
• Familiarity with GPU environments and the relationship between models, GPU compute, and VRAM.
• Understanding of how model size, precision, quantization, context length, batching, and concurrency affect infrastructure requirements.
• Ability to estimate and evaluate infrastructure requirements for different LLM and AI workloads.
• Familiarity with GPU drivers, runtimes, containers, passthrough, and monitoring concepts.
• Ability to investigate GPU utilization and inference-performance issues.

AI / LLM Engineering

• Strong practical understanding of LLMs, Generative AI, RAG, embeddings, vector search, tool calling, AI agents, context management, prompts and AI workflows.
• Hands-on experience building and deploying real AI-powered applications or systems.
• Experience working with commercial LLM APIs, AI development platforms such as Lovable, Replit, Cursor, Bolt, Make, Zapier, n8n, LangChain, OpenAI API, Claude API or equivalent platforms.
• Experience with open-source LLM ecosystems and a practical understanding of self-hosted model deployment.
• Understanding of LLM evaluation, hallucination management, structured output validation, latency, reliability, and cost optimization.
• Ability to decide when an LLM is appropriate and when deterministic software or automation is the better engineering solution.

Infrastructure & Systems

• Strong Linux fundamentals, preferably Ubuntu/Debian or comparable distributions.
• Hands-on experience with virtualization and/or containerized environments.
• Practical knowledge of Docker and modern deployment practices.
• Experience with Proxmox, KVM, VMware, Hyper-V, Kubernetes, or comparable infrastructure technologies.
• Understanding of networking fundamentals including DNS, TCP/IP, firewalls, ports, routing, and service connectivity.
• Experience designing, deploying, debugging, or operating production infrastructure.
• Good understanding of SaaS platforms, hosting businesses, customer portals, admin panels, CRM systems, or sales funnels.

Software Development & Integration

• Strong scripting or programming experience with Python and/or JavaScript/TypeScript/Node.js.
• Experience integrating APIs, webhooks, databases, backend services, and third-party systems.
• Familiarity with PostgreSQL, MySQL/MariaDB, Redis, vector databases, or comparable technologies.
• Experience designing reliable automation and backend workflows.
• Working knowledge of Git and modern software development/deployment practices.
• Ability to create clean UI/UX designs or work closely with designers to produce modern, high-quality interfaces.

Production Engineering

• Experience with monitoring, centralized logging, alerting, health checks, and production troubleshooting.
• Understanding of high availability, failure recovery, backups, and disaster recovery principles.
• Strong debugging and root-cause-analysis skills.
• Security mindset with an understanding of least privilege, credential management, isolation, and controlled system access.
• Good communication skills and ability to explain AI ideas to technical and non-technical teams.
• Ability to research, test, compare, and implement new AI tools quickly.

Professional Requirements

• Strong analytical and problem-solving ability.
• Ability to take ownership of a technical problem from initial investigation through architecture, implementation, testing, and production deployment.
• Ability to communicate technical decisions clearly with both technical and non-technical stakeholders.
• Professional working proficiency in English.
• Ability to work independently in a fully remote environment.
• At least 5 years of hands-on experience in one or more of the following areas:infrastructure engineering, platform engineering, DevOps, site reliability engineering, cloud engineering, backend systems, virtualization, or production systems administration.
• At least 2–3 years of practical experience building, deploying, operating, or integrating AI-powered applications, LLM systems, or AI infrastructure.
• Demonstrable experience taking technical solutions from initial concept or prototype into a production environment.
• Experience owning technical problems across architecture, implementation, deployment, monitoring, and troubleshooting.

Preferred Skills
Experience in several of the following areas will be considered a strong advantage:

• Hosting, cloud infrastructure, datacenter, VPS, dedicated-server, or SaaS environments.
• Proxmox VE and Proxmox Backup Server.
• Proxmox clustering, templates, backups, API automation, and operational troubleshooting.
• KVM virtualization and PCIe/GPU passthrough.
• ZFS, Ceph, shared storage, or distributed storage environments.
• NVIDIA GPU infrastructure and CUDA ecosystem familiarity.
• vLLM, Ollama, LiteLLM, Text Generation Inference, NVIDIA Triton, or similar technologies.
• LLM quantization approaches such as AWQ, GPTQ, GGUF, or comparable methods.
• LoRA, adapter-based fine-tuning, or model customization.
• Multi-GPU inference or distributed AI workloads.
• Model gateways and multi-model routing.
• LangChain, LangGraph, LlamaIndex, CrewAI, AutoGen, or comparable orchestration frameworks.
• Model Context Protocol and tool-based AI integrations.
• Qdrant, pgvector, Weaviate, Milvus, Pinecone, or similar vector technologies.
• Terraform and Ansible.
• Kubernetes or container-orchestration platforms.
• GitLab CI/CD, GitHub Actions, or comparable CI/CD systems.
• n8n, Make, or comparable workflow-automation platforms.
• Redis, message queues, and asynchronous job-processing systems.
• Prometheus, Grafana, OpenTelemetry, Sentry, or comparable observability platforms.
• Vault or comparable secrets-management platforms.
• SSO, identity, and access-management systems.
• WHMCS, cPanel/WHM, HostBill, Blesta, customer portals, billing systems, or hosting automation.
• Incident response, infrastructure hardening, and production-security practices.
• Building internal developer platforms or self-service infrastructure.
• Creating customer-facing infrastructure or platform products.
• Experience exposing AI capabilities through reusable internal or public APIs.

Example Projects This Person May Build

• An AI support agent that reads tickets, suggests replies, classifies urgency, and escalates only important cases to employees.
• An AI sales assistant that analyzes leads, recommends follow-ups, creates offers, and helps the team close more customers.
• An internal AI dashboard that shows company performance, sales trends, support load, churn risk, and improvement suggestions.
• A smart customer onboarding flow that automatically recommends hosting plans, add-ons, upgrades, and setup steps.
• AI scripts that detect repetitive employee tasks and automate them using APIs, webhooks, and scheduled jobs.
• AI-generated landing pages, promotional pages, product explanations, and marketing content for new company services.
• An internal company copilot connected to documentation, SOPs, product data, and support knowledge base.
• AI quality-control tools that review employee replies, sales calls, support tickets, website content, or product descriptions.

Candidate Profile
The ideal candidate is not only an AI user, but an AI builder and systems thinker. They should be able to look at the whole company, understand where time and money are being wasted, and create practical AI solutions that improve performance. They should have strong technical ability, good product thinking, good design taste, and the confidence to promote AI adoption inside the company. They must be able to speak with management, developers, sales, marketing, and support teams, then turn ideas into working tools. They are not someone who sees AI only as prompts, chatbots, or external APIs. You understand that reliable AI applications depend on infrastructure, software architecture, data, security, observability, and operational discipline.

They are comfortable moving between layers. One day, they may benchmark an open-source LLM on GPU infrastructure. Another day, they may configure a containerized inference service, design a RAG pipeline, integrate an agent with an internal API, investigate an infrastructure bottleneck, or automate a repetitive business workflow.

They understand both AI and the systems that AI runs on. They are curious about new models and technologies, but they do not adopt technology simply because it is new. You evaluate whether it solves a real problem, whether it can operate reliably in production, and whether the technical and economic trade-offs make sense.

They are comfortable working independently, making architecture recommendations, rapidly building prototypes, and then applying the engineering discipline required to turn successful prototypes into production systems.

Success Metrics

• Reduction in manual employee workload.
• Faster customer support response and resolution time.
• Higher sales conversion and upsell rate.
• More automated internal workflows.
• Improved quality of company software and customer experience.
• Successful launch of new AI tools, agents, scripts, or platforms.
• Clear documentation and adoption of AI systems by company teams.