AI Solutions & Automation Engineer
Roselyne June
Building intelligent systems where AI, automation and real-world workflows meet.
I create AI applications, intelligent agents and automated solutions that turn real business needs into working systems.
Production AI agent, live
Zawadi is embedded on a real ecommerce storefront and connected to operational support workflows.
AI product in active testing
Job Signal, an AI job-search intelligence platform, is in MVP testing and development.
Automation engineering background
Production automation, integrations and operational workflows across live infrastructure.
Capabilities
What I bring to a build.
A practical mix of AI product thinking, automation engineering, system integration and product delivery.
- 01
AI product & agents
AI applications, agent workflows, LLM integration, prompt and workflow design, structured outputs and local AI experimentation.
- 02
Automation & integration
Python, n8n, APIs, Shopify integrations, workflow orchestration, notifications, escalation and logging.
- 03
Engineering foundation
Infrastructure automation, monitoring, systems integration, NetBox, Nagios and production operational workflows.
- 04
Product & delivery
PRDs, workflow definition, UX direction, AI-assisted development, testing, iteration and deployment.
Selected work
Four systems, four different kinds of proof.
01
Job Signal
Turns a job seeker's own application history into evidence about what is working, what is not, and what to test next.
- Product strategy
- UX direction
- AI workflow
- MVP development
02
Zawadi AI Support Agent
A live support agent connecting customer conversations with Shopify order lookup, escalation and support logging.
- Customer message
- Support agent + Claude
- Order lookup · escalation · logging
03
LocalOps AI
A local-LLM decision prototype that grounds incidents against policy and returns structured decisions for deterministic Python logic.
- 01
Incident input
- 02
Python application
- 03
Policy context + local LLM
- 04
Structured decision
- 05
Deterministic logic
- 06
Action or escalation
Local decision pipeline
- 01
Incident input
- 02
Python application
- 03
Policy context + local LLM
- 04
Structured decision
- 05
Deterministic logic
- 06
Action or escalation
The model interprets supplied policy context; deterministic Python logic owns the final action.
Sample incidents are fictional. No company names, hosts or addresses appear in any diagram.
04
AI Influencer & UGC
A finished 9:16 UGC workflow spanning concept, script, AI visuals, presenter direction and final production.
What I build
Systems that keep going after the answer.
Four areas of work, from first workflow sketch to a system running in production.
- 01
AI applications & product development
Define the workflow and UX, then build and test a functional AI-enabled MVP.
- 02
AI agents & customer support systems
Agents that connect to real business data, follow bounded workflows and hand off when needed.
- 03
Business workflow automation
APIs, systems and operational processes connected into reliable, observable workflows.
- 04
AI-enabled websites & digital experiences
Polished digital experiences built around a clear business goal, with AI features and workflow integrations where they add real value.
Engineering foundation
Built beyond the prototype.
Before AI products, I built and ran production automation: validation, notification, execution, verification and escalation across live infrastructure.
- locations supported through automation
- 50+locations supported through automation
- VMs reconciled through custom infrastructure tooling
- 400+VMs reconciled through custom infrastructure tooling
- monitored devices synchronized across monitoring systems
- 950+monitored devices synchronized across monitoring systems
About
From infrastructure automation to AI products.
I moved from automation and infrastructure engineering into AI applications, agents and product development — and I still care most about what happens after the model produces an answer.
Python and operational tooling, systems integration, event-driven workflows: the same engineering discipline now shapes how I design AI systems that businesses can actually run.
Working areas
- AI & intelligence — applications, agents, grounding, evaluation
- Build & automate — APIs, integrations, workflow automation
- Engineering — Python tooling, systems integration, monitoring
POD 7
We turn AI ideas into working tools that people actually use.
POD 7 & AfriKonnect
“We turn AI ideas into working tools that people actually use.”
A collaborative product developed by POD 7, from team ideation and GPT-assisted PRD development through to a working platform built in Emergent.
Work with me
Have an idea worth building?
If you’re exploring an AI application, agent, automation or digital product, tell me what the system needs to do.