Enterprise AI agents · edge + hybrid cloud

Move one governed AI workflow into production.

Geek Nexus helps enterprise teams connect models to private data, existing systems, approvals, and day-to-day operations through one repeatable deployment path.

Email our team Explore solutions

Private by design Connected to existing systems Built to scale

Best fit Mid-sized enterprises with sensitive data One repeatable workflow to improve

Designed for CTOs, CIOs, digital-transformation leaders, and operations owners who need measurable value without giving up governance.

Meet the founder

Connecting enterprise needs with practical delivery.

Celia Shu connects enterprise needs, technology resources and commercial execution. Since 2017, her work in enterprise cloud and edge computing has covered needs discovery, solution coordination, negotiation and cross-functional project development.

Earlier procurement roles in aviation manufacturing and industrial technology involved materials supply, contracts and ERP workflows. That operational background helps her understand the processes and constraints a technology project must fit.

2012-2017

Procurement & supply chains

Procurement roles at AVIC South China and Schneider Electric involved materials supply, contracts and enterprise purchasing workflows.

Since 2017

Enterprise cloud & edge

Roles at Wangsu, Volcano Engine, Baidu AI Cloud, Tencent Cloud and UCloud involved customer discovery, solution coordination and commercial development.

Since 2020

AI applications & ecosystem collaboration

Experience includes speech, translation, customer service and outbound-calling projects, alongside coordination of cloud and industry resources.

Experience from previous roles

These anonymized examples describe projects Celia participated in or advanced while working for previous employers. The engagements belong to those employers and are not presented as GeekNexus customer results or current partnerships.

AI service workflows

Customer service & speech applications

Business need: Enterprises needed speech, conversational or outbound-calling capabilities connected to existing service workflows.
Celia’s role: At Baidu AI Cloud and Volcano Engine, Celia contributed to opportunity discovery, solution-resource coordination and commercial development.
Project experience: Her prior experience includes integrated customer-service solutions and AI outbound-calling projects brought into use.
Applied today: Define the service workflow, system interfaces and human handoff before selecting the technology.

Hybrid cloud & cross-border

Cloud coordination for gaming

Business need: Domestic and overseas gaming operations needed coordinated cloud, acceleration, security and deployment resources.
Celia’s role: At UCloud and Tencent Cloud, Celia advanced gaming-sector cloud engagements and coordinated internal and external solution resources.
Project experience: Past projects involved hybrid cloud, CDN, security and cross-border acceleration.
Applied today: Agree deployment regions, performance needs, supplier responsibilities and delivery scope early.

Retail & connected hardware

Connecting operations to cloud services

Business need: Retail and connected-device businesses needed networks, security, devices and cloud services to work together.
Celia’s role: At Wangsu, Baidu AI Cloud and other employers, Celia contributed to customer development, needs alignment and solution partnerships.
Project experience: Experience includes retail acceleration and security projects, plus IoT and data-solution adoption for connected hardware.
Applied today: Start with devices, data and existing systems to identify the right cloud and edge collaboration model.

Solutions

Three practical places to start.

Each solution begins with one bounded workflow and includes the data, integration, governance, and operating requirements needed for real use.

01KNOWLEDGE

Private enterprise knowledge agent

Governed retrieval across internal documents
Role-aware answers with source traceability
Private or hybrid execution options

For teams that need reliable internal answers without moving sensitive knowledge into an uncontrolled public workflow.

02SERVICE

Customer and employee service agent

CRM, ticketing, knowledge, and API integration
Approvals, escalation, and human handoff
Quality, latency, and usage monitoring

For service teams that want faster resolution while preserving accountability and a clear path to human support.

03OPERATIONS

Document and workflow automation

Document intake, extraction, and validation
Traceable multi-step agent workflows
Exception handling and human checkpoints

For operations teams that need to reduce repetitive professional work without creating an ungoverned black box.

Ideal customer profile

Built for teams with a real workflow to improve.

The strongest starting point is not a broad AI ambition. It is a named business owner, a bounded workflow, and a result the team can measure.

Sensitive or regulated dataRepeatable, high-volume workflowExisting knowledge base, CRM, ERP, or APIsNamed business and technical owner

CONTROLGoverned data and access

Define where data lives, who can use it, and how every action is traced.

VALUEMeasures agreed before build

Set quality, time, cost, adoption, and risk targets before the pilot begins.

ADOPTIONHuman review and operations

Design approvals, exceptions, monitoring, and ownership into the workflow.

Start with one practical workflow

See whether AgentBox fits.

Bring one workflow, its users, and the systems it touches. We will help frame the data boundary, pilot scope, and measures for a go-or-no-go decision.

Email[email protected] Phone (China)+86 166 0214 6315

LocationShanghai, China