AI ML Architect / Senior FDE Lead

A senior technical leadershiprole at the intersection of architecture rigor, hands-on engineering, andproduction-grade delivery.

Anthrobyte builds production-grade enterprise AI systems forglobal clients across industries. As we build out our presence in the U.S. fromSan Francisco, we need a senior AI/ML architect who can hold the full technicaljourney: system architecture, agentic and LLM infrastructure design, hands-onbuild, and production deployment — all as one coherent capability.

This is not a role where you hand off an architecture diagramand move on to the next project. You will work directly with the CTO, foundingteam, and enterprise clients to define how AI systems are architected,engineered, and hardened for production inside complex organisations. You willbe the person who walks into a room with a messy, high-stakes technical problemand walks out with an architecture the team can actually build — and thenbuilds it.

Requirements

Own the architecture. Ship itto production.

·    Lead architecture discovery with enterpriseclients — assess data maturity, system landscape, and AI readiness before asingle line of code is written

·    Own end-to-end system design for AI proposals:model selection, RAG/agentic architecture, data pipeline design, infrastructuresizing, and ROI framing

·    Build working prototypes, architectureblueprints, and technical proof points that de-risk the engagement before fullbuild

·    Set technical standards and reusablearchitecture patterns that the broader engineering team builds on

·    Be hands-on in the build — write productioncode, design data pipelines, and stand up the infrastructure your architecturecalls for

·    Own deployment realities: integrationcomplexity, security and compliance constraints, scaling, and observability

·    Drive go-live milestones, uptime, andpost-deployment performance with direct accountability for outcomes

·    Debug, iterate, and adapt in production — whensomething breaks, you own the fix, not just the postmortem

·    Act as the trusted technical authority to clientstakeholders — CTOs, VPs of Engineering, platform teams — not just a vendor onthe call

·    Translate deep technical tradeoffs into languagebusiness stakeholders can act on, without losing the substance

·    Collaborate with presales and delivery teams toensure every commitment made to a client is technically buildable

·    Build long-term technical trust with clientsthat turns single engagements into expanded, multi-year mandates

·    Define AI architecture patterns for the firm:LLM orchestration, RAG pipelines, agentic workflows, and evaluation frameworks

·    Stay ahead of applied AI research and emergingframeworks; bring what matters back to the team before it’s common knowledge

·    Mentor AI engineers on both architecture rigorand forward-deployed delivery craft

·    Contribute to Anthrobyte's technical knowledgecapital — architecture playbooks, reusable accelerators, and internal tooling

You operate across the full stack — from distributed systemsdesign to client boardroom. You simplify complex architectures without losingwhat makes them robust, and you're as comfortable in a production incidentchannel as you are presenting system design to a VP of Engineering.

·    8+ years of hands-on AI/ML and softwarearchitecture experience, with production systems at real scale

·    Proven track record architecting and shippingLLM-based or agentic AI systems into production — not just PoCs

·    Deep hands-on fluency: Python, distributedsystems, LLM frameworks (LangChain, LlamaIndex, HuggingFace), and cloudinfrastructure (AWS, Azure, GCP)

·    Experience designing end-to-end AI/ML platforms:data pipelines, model serving, evaluation, and monitoring

·    Strong client-facing communication — able todefend architecture decisions to both engineers and executives

·   Comfort operating in ambiguity and 0→1 environments; startup orforward-deployed experience strongly preferred

·   A bias toward ownership — you close loopswithout being asked, and you treat production issues as yours to fix