Skip to content
Currently available — for the right work·France903+ Day French Streak·2026 Q2 calendar — open now
AMDamd·fauzulThe job description
Silo AI · Principal Applications Eng · Vancouver

Build the apps thatdeploy, run and scaleAI on modern GPU infrastructure.

AMD Silo AI is hiring a Principal Software Development Engineer for enterprise applications and platform engineering — owning data-heavy backends and user-facing apps for AI/GPU workflows, setting Kubernetes toolchain and API direction, and driving AI-first development practices. I bring a decade of player-coach delivery, production agentic systems, and customer-facing platform work — and I am ready for Vancouver hybrid.

Ask Fauzul
9+ yrs
hands-on principal / founder delivery
4×
faster time-to-market via CI/CD discipline
9
client geographies shipped into

Enterprise AI apps.Platform discipline.Customer-adjacent ownership.

Silo AI sits where GPU infrastructure meets the tools customers actually use to deploy and scale ML/LLM workloads. The principal seat is not only code — it is architecture under load, roadmap influence with customers, and raising engineering excellence across teams. That is the work I have practised as CIO and player-coach: ship the system, set the standards, mentor the engineers, feed field learning back into the product loop.

Accelerated · Substantial · Trusted
01
Accelerated

Ship AI infrastructure apps that move under real load

I design and deliver data-heavy backends and React/TypeScript UIs for cloud-deployed platforms — from HIPAA clinical systems to agentic production pipelines — with CI/CD, observability and API standards as defaults, not afterthoughts.

02
Substantial

Architecture that survives customers and scale

Principal work means setting direction for toolchains, APIs and integrations while debugging cross-cutting production issues. At I owned that loop end-to-end — feasibility with product, phased delivery, and measurable milestones across nine geographies.

03
Compounding

Engineers grown into leaders, not just features shipped

Weekly design and code review, 1:1 coaching, judgment over throughput — alumni I mentored now lead engineering in Norway (2), Canada and Germany. That is the same bar-raising a principal seat needs to clear: not just your own code, the team's.

Together we advance AI infrastructure — by building the applications customers actually run on GPU and cloud platforms, with engineering excellence that holds under load.

Fauzul — on why AMD Silo AI
JD map

Requirements as modular packages.

Mapped to the AMD Silo AI posting — responsibilities and preferred experience, verdict by verdict, with honest adapt paths where GPU/K8s depth is partial.

platform

Platform & Kubernetes

Secure AI infrastructure apps, K8s toolchains, APIs/UIs, and cloud-native pragmatism.

Strong01

Design, build, and evolve secure, reliable AI infrastructure applications and tools that help customers deploy, run and scale ML and LLM workloads on modern cloud and GPU infrastructure

I ship cloud-native, data-intensive applications and AI tooling under real customer load — including a Claude-powered production CMS () that turns plain-English intent into gated deploys, plus agentic systems (VisaPros, NewScriber) that run ML/LLM workflows end-to-end. Security and reliability are product surfaces (HIPAA, GDPR/WCAG across five organisations).

Adaptable02

Own and drive design of complex, data-heavy backend services and user-facing applications for AI/GPU and platform workflows (Kubernetes, APIs, integrations)

Strong on data-heavy backends, APIs and React/TypeScript UIs for platform workflows ( clinical platform; agentic services in Go/Python/TS). Kubernetes / orchestrated environments: familiarity in my stack — Docker, Cloud Run, ECS, CI/CD — with production depth heavier on managed containers than operating GPU-aware K8s clusters day-to-day.

Adapt pathDeepen Kubernetes toolchain ownership (operators, GPU scheduling hooks, cluster-facing APIs) as the first 90-day ramp — the API/UI/service design muscle transfers; the AMD-specific K8s surface is the learn-on-the-job layer.

Adaptable03

Set and evolve architectural direction for Kubernetes toolchain, APIs and UIs enabling core AI workflows; ensure security, scalability, and maintainability under real customer load

I have set architectural direction for APIs, UIs and integrations under customer load, with security/compliance designed in. Full K8s toolchain architecture for AI workflows is adjacent rather than a titled specialty today.

Adapt pathLead first on API/UI architecture and security baselines, while pairing with Silo AI platform specialists on GPU-aware K8s internals until that surface is second nature.

Adaptable04

Open-source or platform-ecosystem awareness (K8s operators, schedulers, observability stacks)

Solid observability and platform-ecosystem awareness from cloud-native delivery; K8s operators/schedulers are known concepts with less hands-on operator authorship.

Adapt pathContribute to operator/scheduler-facing work after orientation — start by hardening observability and API contracts the ecosystem already expects.

Strong05

Hands-on experience with relational and NoSQL databases, Git, and Kubernetes (or equivalent orchestrated environments)

PostgreSQL and Git are deep muscle. NoSQL is not a gap: MongoDB, MongoDB Atlas, DocumentDB, and DynamoDB across a run of document-store projects since 2017. Orchestration: strong on Docker, Cloud Run, ECS; Kubernetes listed as familiarity — equivalent orchestrated environments yes, deep production K8s day-job no.

Adapt pathProduction K8s (operators, schedulers, cluster-facing tooling) deepens on the job; relational, NoSQL, Git, and container orchestration already transfer.

ai

AI, Agents & Performance

Agent-driven engineering, GPU/LLM-adjacent systems, and production reliability under load.

Strong01

Optimize systems for efficiency, performance and reliability; debug and resolve cross-cutting production issues

Production debugging across distributed systems is routine — from HIPAA clinical platforms to multi-agent pipelines. Drove 4× faster time-to-market (75% cycle-time reduction) via repeatable workflows, cloud cost discipline and consistent delivery mechanisms.

Adaptable02

Stay current on cloud-native, Kubernetes, GPU/ML workflows, and LLM infrastructure trends and apply pragmatic choices to our stack and practices

Current and pragmatic on cloud-native, LLM infrastructure and agentic workflows (daily). GPU/ML training-cluster internals and AMD-specific hardware stacks are lighter — I follow the trends and have shipped ML/LLM apps, not GPU kernel or ROCm platform code.

Adapt pathTreat AMD GPU / Instinct / ROCm-adjacent platform literacy as a deliberate study track in the first quarter, grounded by pairing with Silo AI GPU specialists while owning the application layer.

Strong03

Drive AI-first development practices and continuously promote engineering efficiency using modern agent-driven software development methods

Claude-led daily development (Gemini second — model and harness picked per task, not vendor loyalty), Anthropic Claude Code 101 and Introduction to Agent Skills certifications, production agent systems ( A2A, , CMS). Promoting agent-driven methods is already how I work and coach.

Adaptable04

Experience with GPU workloads, ML/LLM training or inference pipelines, or multi-tenant platform concerns

Strong on ML/LLM application and inference-adjacent pipelines (agent orchestration, evals, production LLM apps). GPU training workloads and multi-tenant GPU platforms are not a titled specialty — WebGPU on-device for this site's chat is adjacent curiosity, not Instinct-class ops.

Adapt pathLead on the application/control-plane side of ML/LLM workflows immediately; ramp on GPU workload and multi-tenant scheduling concerns with Silo AI platform peers — named honestly, not overclaimed.

leadership

Leadership & Mentorship

Principal influence across teams — architecture calls, coaching, and growing judgment.

Strong01

Combine hands-on full-stack engineering with technical and strategic leadership: influence roadmap and architecture, partner with product and customers, raise the bar for quality, operability, and developer experience

This is my default seat: co-founder → Engineering Manager → CIO at , still hands-on daily, influencing roadmap with product/customers while embedding CI/CD, review and observability so quality is a habit. Mentored engineers who now lead in Norway, Canada and Germany.

Strong02

Lead technically and strategically across teams and drive complex issues (technical and organizational) to resolution

Cross-team technical leadership as CIO of a 22-person cross-functional core (design, engineering, QA, product, business; 52+ with contractors) — resolving both code and organisational blockers.

Strong03

Mentor and uplift engineers via design reviews, code reviews, and coaching to grow technical judgment and execution quality of the team

Weekly 1:1s and review culture as default. Alumni now lead in Norway (2), Canada and Germany, plus two leading local firms — mentoring is a measured outcome, not a soft claim.

Strong04

Prior technical leadership: architecture decisions, cross-team initiatives, and growing engineers

Architecture decisions and cross-team initiatives as EM/CIO; growing engineers into leadership abroad is documented.

delivery

Delivery, Customer & Roadmap

CI/CD and observability, enterprise AI direction, field loops, and cross-function trade-offs.

Strong01

Embed engineering excellence: code quality, automated testing, CI/CD, observability, and API/design standards across the teams you work with

Made TDD, review and CI/CD non-negotiable at ; agent-authored changes at only ship through verification gates. Observability and API standards are how I keep multi-geo delivery predictable.

Strong02

Influence Enterprise AI roadmap through deep understanding of customer needs, market direction, and technical feasibility; translating pain into prioritized engineering work

Decade of customer discovery and feasibility under founder titles — translating fuzzy enterprise briefs into phased delivery across nine geographies. CSPO discipline for prioritisation; product partnership is not a side skill.

Strong03

Interface with customers (and sales/customer success as needed): clarify requirements, manage expectations, and feed learnings back into the product/engineering loop

Field-adjacent by default: 4 years embedded with Skytracks Canada's technical CTO ('s Angular/WebSocket collaboration), Norway Frontgo pod 0→production, and ongoing client technical partnership. Feedback into the engineering loop is how stayed delivery-credible.

Strong04

Lead initiatives from evolving requirements to shipped outcomes—decompose large problems into phases, owners, and measurable milestones

0→1 across five verticals and multi-quarter enterprise programmes — phased ownership, milestones and shipped outcomes are the operating system, not a slide.

Strong05

Represent engineering in roadmap and trade-off discussions with product, design, and platform, and align delivery with business and UX goals

UX-engineer origins plus founder/CIO seat — I speak product, design and engineering without a translation layer, and I still design in Figma when craft matters.

Strong06

Ability to work across functions (product, design, platform, GTM) with clear communication and stakeholder management

Cross-functional by construction at and in client pods — product, design, platform and GTM stakeholders are the default room.

Strong07

Agile delivery experience; comfort owning end-to-end design → implementation → deployment → support

Agile delivery end-to-end for a decade, including post-launch support and escalation ownership under founder titles. CSPO.

Strong08

Initiative, ownership, and collaboration; pragmatic problem-solving under constraints

Bootstrapped studio constraints, multi-geo clients, compliance-heavy domains — pragmatic ownership is the default mode.

Strong09

Customer-facing or field-adjacent engineering: discovery, technical scoping, or success of complex B2B deployments

Complex B2B deployments across nine geographies — discovery, scoping and success under founder/strategic-partner titles (Frontgo, , , ).

foundations

Foundations

Principal track record, Python/TypeScript/React depth, education bridge, Vancouver hybrid.

Strong01

Proven track record as a senior/principal-level full-stack or backend-leaning software engineer in cloud-deployed, data-intensive web applications

Decade of senior/principal-equivalent delivery: cloud-deployed, data-intensive apps across health-tech, music-tech, hospitality and agentic products — still hands-on as CIO.

Strong02

Strong Python and modern JavaScript/TypeScript and a major UI framework (React or equivalent); solid API and service design

TypeScript/React daily; Python for AI/agent work; solid API and service design across Node, Go and FastAPI-style services ( country agents, ConnectRPC, Go orchestration).

Strong03

Location: Vancouver, British Columbia — hybrid

Canada is not just a target geography — 4 years of direct client delivery ( & SkyTracks, Angular/WebSocket collaboration with the client's technical CTO), fractional CTO work in Quebec, FDE engagements in Toronto, and six months already lived in Toronto, on top of Global Talent Stream relocation intent. Ready for Vancouver hybrid.

Adaptable04

Bachelor's or Master's in Computer Science, Engineering, or equivalent experience

Field of Study: Computer Science & Engineering · North South University (2012–2015) — then straight into building. Equivalent experience: decade of shipped principal-level work.

Adapt pathCredential is experience-based rather than a completed degree; the JD allows equivalent experience, which is the operative qualification.

DailyClaude-led coding with Gemini when the task fits — agent skills, hooks and evals as habit
Every initiativeDecompose large platform problems into phases, owners and measurable milestones
Every releaseCI/CD, observability and API standards enforced — not optional
Every reviewMentor via design and code review — grow judgment, not just throughput

Platform apps. Agent-driven delivery. Customer load.

I build the applications and tooling that let teams deploy AI safely — then I keep raising the bar on quality, operability and how the team ships with agents.

Global Jute — Claude-powered CMS behind CI/CD gates

A production natural-language CMS for a 26-year export house: plain-English intent becomes agent-authored changes that only reach production through verification gates. It is the clearest proof of the AI-first engineering discipline Silo AI's mandate calls for — secure, gated, operator-trusted delivery — built Claude-led, with engineering excellence as the product surface.

  1. 01

    Plain-English requests → agent → PR → automated deploy, no developer in the critical path

  2. 02

    CI/CD verification gates as the trust boundary for agent-authored changes

  3. 03

    Customer-facing delivery inside a live commodity-export business, not a demo

How I operate

Principal cadence for Silo AI apps.

How I would operate inside Silo AI Enterprise applications — the same loop I already run on agentic and platform work.

  1. 01

    Own the hard path in the code

    Principal means hands-on on the gnarly slice — API design, data models, UI workflows for AI ops — not architecture theatre. I still ship daily in TypeScript, Python and Go.

  2. 02

    Make excellence automatic

    Embed testing, CI/CD, observability and design standards so quality does not depend on heroics. That discipline is what cut cycle time 4× at .

  3. 03

    Drive AI-first development for real

    Daily Claude-led engineering (Gemini second), agent skills, hooks and eval harnesses — the exact modern agent-driven methods the JD asks to promote, already in production on client and portfolio systems.

  4. 04

    Close the customer loop

    Clarify requirements with customers and GTM partners, manage expectations, and turn field pain into prioritized engineering work — the motion I have run under founder titles for a decade.

Top strengths

Lead signal

Principal full-stack + leadership in one seat

Co-founder/CIO who still ships TypeScript, Python and Go — roadmap influence and code review in the same week.

AI-first, agent-driven development in production

Claude-led stack, Anthropic certifications, CMS + + — the JD's AI-first mandate is already how I work.

Customer-adjacent B2B delivery

Nine geographies, compliance-heavy platforms, discovery through support — field learning fed back into engineering.

Engineering excellence as default

4× faster time-to-market via CI/CD and review discipline; agent changes gated by verification, not vibes.

Technical inventory

Stack & receipts.

The stack and receipts behind the claims — Claude-led for production delivery, model and harness picked per task.

Languages & UI
TypeScriptPythonGoReactNode.js
Data & APIs
PostgreSQLREST/gRPC-style APIsConnectRPCevent-driven patterns
Cloud & delivery
DockerGitHub ActionsGoogle Cloud RunAWS ECSCI/CDKubernetes (familiarity)
AI & agents
Claude CodeAgent SkillsMCPA2AGeminieval harnessesn8n
Introduction to Agent SkillsClaude Code 101Certified Scrum Product Owner® (CSPO)

Named ramps

Honest bridges — not buried. Read left to right like a bring-up schedule.

First 90 days

Deep Kubernetes toolchain ownership for GPU-aware AI workflows

Transfer API/UI/service design strength; pair on operators, scheduling and cluster-facing tooling until K8s is a daily surface.

First quarter

AMD GPU / multi-tenant GPU platform internals

Own the application and control-plane layer immediately; build Instinct/ROCm-adjacent literacy with Silo AI GPU specialists — never invent hardware credentials.

Ongoing on the job

K8s operators/schedulers authorship

PostgreSQL, NoSQL (MongoDB/DocumentDB/Atlas), Git, and container orchestration all transfer; deepen operator and scheduler authorship as the platform demands.

Hard questions

The objections a sharp principal hiring loop would raise — answered straight.

Not as a titled GPU-platform engineer. I have shipped ML/LLM applications and agentic inference-adjacent pipelines under real customer load, and I am candid that multi-tenant GPU scheduling and AMD hardware internals are a ramp — I would own the application layer from day one while building that literacy with Silo AI specialists.

Production strength is in cloud-native orchestration — Docker, Cloud Run, ECS, CI/CD — with Kubernetes as familiarity rather than a years-long day job. The principal skills the role needs (API/UI architecture, data-heavy services, standards, customer loop) transfer; deep K8s toolchain ownership is the named 90-day ramp.

The JD asks for AI-first, agent-driven development methods — that discipline is model-agnostic. My production AI has been Claude-led with Gemini second, and I pick the model and harness that fits each task rather than defaulting to one vendor — Claude leads client delivery today because it earns that seat. I already promote agent skills, hooks and evals as engineering practice.

Yes — and it isn't a cold bet. I've delivered directly for Canadian clients for four years ( & SkyTracks, working with the client's technical CTO), plus fractional CTO work in Quebec and FDE engagements in Toronto — and a family visit gave me six months in Toronto firsthand, so the country isn't unfamiliar. This Silo AI hybrid seat is the kind of principal platform work I'm moving for, on top of real professional history with Canadian teams.

Field of Study: Computer Science & Engineering at North South University (2012–2015). I do not claim a completed degree. The JD allows equivalent experience — a decade of principal-level shipped systems is the operative credential.

Beyond the checklist

What I bring past the JD.

The posting maps the principal bar. These are the extras that show up once the hire is in the room.

Founder operator

Roadmap influence from real constraints

Co-founder → EM → CIO at : I have sat in the trade-off room with product, design and GTM, and owned outcomes when the brief was fuzzy.

Agent-native

AI-first development already in production

(six-agent A2A on Gemini), (Go + n8n + Kimi via OpenRouter + Gemini TTS via Azure + evals), CMS — plus Anthropic certifications as dated proof, not slideware.

Security-minded

Compliance treated as a product surface, not a tax

HIPAA, GDPR and WCAG baked in from day one across five client organisations — the same instinct behind the JD's own 'secure, reliable' mandate, headed off as design work rather than a retrofit.

Consistency

900+-day public learning streak

The same compounding discipline I bring to shipping cadence and engineering standards — visible, not claimed.

Public consistency · 982-day French streak

Principal Applications Eng · Silo AI

together we advance_AI applications that scale.

Canada is not a fresh bet: four years delivering for AudioBin & SkyTracks (Angular/WebSocket collaboration with the client's technical CTO), fractional CTO work in Quebec, FDE engagements in Toronto, and six months already lived in Toronto. Vancouver hybrid for this Silo AI seat is deliberate relocation intent on top of real Canadian client history and lived experience in the country. See Vancouver for the full picture.

Email Fauzul Book 30 min

together we advance_

Application fit for Principal Software Applications Eng at AMD · Official JD · AMD hub ·