Principal Forward Deployed Architect, Gemini Enterprise Platform (GCP)

Role
Summary 

You are the
person a client trusts to turn an ambitious Gemini Enterprise vision into a
business outcome that lasts. As Principal Forward Deployed Architect on an
account, you own the result -- the value the client set out to create — and
with it the technical whole that produces that value: the GenAI platform
foundation on Google Cloud, the agent landscape built on the Gemini Enterprise
Agent Platform (GEAP), the context-graph and data foundation those agents
reason over, and the enterprise rollout into the Gemini Enterprise app. 

Where
specialist engineers each own an individual agent, MCP server or data pipeline,
you own the whole — deep in the agent platform and the context / data
foundation, fluent enough across governance, runtime and adoption to design,
sequence and defend the program end to end. You are the senior technical
counterpart the client's executives call before they have decided what to
build; more importantly, you are the reason they keep calling. You make Google
Cloud's AI foundation deliver the outcomes. 

This role
exists because standing up a production agent ecosystem on GEAP is not a
single-layer problem — model choice, agents, grounding graph, governance
perimeter and change management are load-bearing on one another — and because
our largest clients will accept only one senior technical owner rather than
several. 

Deployment
Model 

Placed at
one large Gemini Enterprise account, or holding technical ownership across two
or three smaller concurrent engagements. You may direct AuxoAI delivery teams,
including offshore and onshore Forward Deployed Engineers and client engineers,
on the same program — you own the design coherence across it. Significant
pre-sales involvement is expected: the GEAP target architecture, the GCP
landing-zone approach, effort estimates, and the technical case in proposals
for the practice's largest Gemini opportunities. 

Key
Responsibilities 

Whole-program
architecture 

• Own the target architecture
  across four layers — GCP GenAI platform foundation, the GEAP agent
  landscape, the context-graph / data foundation, and enterprise adoption —
  and sequence delivery across all four. 
  

• Set the reference patterns for
  how agents are built (ground-up in ADK vs. forked and hardened from Agent
  Garden templates), where they run (Agent Engine managed vs. Cloud Run vs.
  self-managed GKE), how they are isolated (sandbox strategy), and how they
  are governed. 
  

• Design the context-graph
  foundation — BigQuery, BigQuery graph (GQL) and/or Spanner Graph — and the
  grounding / RAG strategy that connects it to agents, including entity
  resolution, semantic modelling and retrieval over Vertex AI Vector Search. 
  

• Identify decisions in one layer
  that are load-bearing for others (e.g., a grounding-data residency choice
  that constrains the runtime target and the governance perimeter) and force
  them to resolution before delivery commits, not during it. 
  

• Arbitrate cross-track
  trade-offs where multiple Forward Deployed Engineers are deployed to the
  same client, with a written rationale. 
  

• Maintain technical proximity:
  review agent designs and evaluation results, interrogate trajectory and
  latency behaviour, participate in incident reviews, and perform selective
  hands-on work where it materially changes the outcome. 
  

• Represent AuxoAI in the
  client's security, compliance and architecture review boards, including
  the model-governance and data-governance forums. 
  

Client
and commercial 

• Advise client executives on
  trade-offs, sequencing, delivery risk and what not to build — including
  which use cases are not yet safe to automate. 
  

• Own the technical scope,
  estimate and defence of proposals and statements of work for the account
  and for major Gemini Enterprise prospects. 
  

• Give AuxoAI leadership an
  accurate read on delivery risk, including remediation plans and
  consumption-cost exposure (runtime vCPU-hours, Sessions and Memory events,
  model tokens, sandbox compute). 
  

Enablement
and practice contribution 

• Enable the client's own
  platform, data and security leadership to operate and extend the agent
  landscape and context graph, with named client owners for each major
  component. 
  

• Develop the Forward Deployed
  Engineers working alongside you on the account, whether or not they report
  to you. 
  

• Contribute GEAP reference
  architectures, context-graph patterns, estimation models and governance
  blueprints that raise the practice standard. 
  

Outcome
Ownership 

You are
accountable for the outcome, not the artifact. Long after AuxoAI rolls off, the
client's agent ecosystem has to keep earning its place — grounded, governed,
evaluated and adopted, still delivering the business result it was built for.
When an agent delivered under your architecture regresses, leaks data, breaches
a policy or loses the users it was meant to serve, you own the explanation to
the client and the plan to make it right. 

Technical
Environment 

Expert
depth in at least two of the areas below; working competence in all. 

Minimum
Qualifications 

1. Master's degree in Computer
   Science, Engineering, Information Systems or a related field, or
   equivalent practical experience. 
   

2. 12+ years in engineering,
   architecture or technical delivery leadership, including senior technical
   ownership of production systems. 
   

3. Expert depth in at least two
   of: the agent / GenAI platform layer, the context-graph and
   semantic-modelling layer, the cloud data-platform layer, and the cloud
   runtime / governance layer — with working competence across the rest,
   demonstrable through an architecture walkthrough. 
   

4. Delivered at least one
   production GenAI or agent system on GCP (Vertex AI / Agent Platform) or a
   directly comparable cloud, including grounding over an enterprise data or
   knowledge foundation. 
   

5. End-to-end ownership of
   technical design for at least two client engagements or major cross-team
   programs, from discovery through production. 
   

6. Experience as the single senior
   technical counterpart to a client's executive team on an engagement of
   material size. 
   

7. Hands-on depth in BigQuery and
   at least one graph or semantic store (Spanner Graph, BigQuery graph, Neo4j
   or equivalent). 
   

8. Delivery inside at least one
   regulated environment, with the ability to describe a design decision the
   regulation forced. 
   

9. Experience owning the technical
   scope, estimate and defence of a proposal or statement of work. 
   

Preferred
Qualifications 

• Hands-on with the Gemini
  Enterprise Agent Platform specifically — ADK, Agent Garden, Model Garden,
  Agent Engine — or a rapid, demonstrable path to it from adjacent agent
  frameworks (LangGraph, CrewAI, Amazon Bedrock Agents, Azure AI Foundry). 
  

• Experience designing and
  operating MCP servers (off-the-shelf, third-party and custom) and
  multi-agent (A2A) topologies. 
  

• Experience building a knowledge
  / context graph for retrieval grounding at enterprise scale. 
  

• Google Cloud Professional
  certification (Cloud Architect, Machine Learning Engineer, or Data
  Engineer). 
  

• Consulting, systems-integrator
  or professional-services background at principal or equivalent level. 
  

• A record of developing senior
  engineers or architects, and of leading hybrid onshore / offshore teams at
  scale. 
  

• Experience deciding against a
  technically attractive approach for commercial, cost or governance
  reasons, and defending that to both client and internal stakeholders.