Skip to content
Franklin Nweke logo

I build websites, apps, and automated workflows that run real operations.

Full-stack engineering across React, Next.js, Node, Python, cloud delivery, and AI-assisted workflows.

Franklin Nweke
Websitesclear public faceAppsworking product layerAutomationsoperations that keep moving
MontechValues.coDerived.artJobaMMG ConciergeAffixdotMonytizeCashboxHerbilityRobes n Stitches

The public face and the operating layer, built together.

The website or app gives the idea a place to live. The engineering behind it handles the repeatable work: data, requests, support, content, handoffs, AI review gates, and the internal tools a team can actually run around.

01

Business Intake & Website Systems

Clear websites, landing pages, forms, and request flows that help customers act and help teams respond without chasing details manually.

02

Custom Web Apps & Internal Tools

Dashboards, portals, request trackers, quoting systems, and admin panels shaped around the way the operation actually runs.

03

AI Workflow Automation

LLM-assisted workflows that classify, draft, route, and check work inside existing tools, with evaluation standards and human approval where the risk calls for it.

Operations trace

Check the server baseline before putting an automation service behind it.

Inspecting host

A short operational trace of how the visible request flow becomes routing, validation, and dependable automation underneath.

8+

Years shipping production systems

15+

Client systems shipped to production

20,000+

Active users across applications

Results, not decks.

Enterprise Web Systems · Montech

Shipping client systems with cleaner deployment flow

Delivered web applications and workflow systems across product, engineering, design, and client stakeholders, with CMS, DAM, API, and automation integrations.

Deployment time reduced by 40% through CI/CD improvements.

ReactNode.jsPythonAzure DevOpsGitHub Actions

LLM Evaluation · Outlier

Catching weak reasoning before AI reaches production

Evaluating model outputs for factual accuracy, instruction following, hallucination risk, and unsupported reasoning.

Practical judgment for safer AI workflows and human-in-the-loop systems.

LLM evaluationRubricsPrompt designGuardrails

Agentic Operations · Agentic-Hub

Turning lead research into a reviewed operating system

Built a local workspace for lead research, qualification notes, CRM tracking, outreach drafts, and human approval before anything gets sent.

A repeatable agentic workflow with provenance, compliance checks, and manual control.

Codexn8nPythonJSONLCRM

Bring the idea. I'll help shape the system.

In a conversation, we can turn the rough version into a real path: the website or app, the tools behind it, and an automation layer that helps the business run.