Skip to main content
HarLyn
Missions & EvidenceVerified Production Architectures

Systems Documented in the Field

We do not publish hypothetical case studies or invented metrics. Every system below is a live or validated production implementation spanning e-commerce AI, RAG retrieval, n8n orchestration, and machine learning network defense.

Verified Client Evidence

Real Systems Running in Production.

Every project demonstrates measurable performance: AI sales platforms, consumer payment engines, and intelligent knowledge assistants.

PROJECT 01 · COMMERCE / FINTECHPublic production system

StarCall Phones — Billboard Device Financing Engine

High-speed smartphone instalment financing platform handling high-volume daily applications driven by citywide billboard campaigns across Nairobi.

Business Outcome:

Processed thousands of consumer financing applications with zero downtime, instant M-Pesa transaction confirmation, and real-time instalment ledgering.

Architecture Stack:
Next.jsTypeScriptTailwind CSSM-Pesa Daraja APIPostgreSQL
Marketing TrafficCitywide Billboards
Platform Speed< 0.8s Mobile Load
Explore Case Study
PROJECT 02 · COMMERCE / AI SALESPublic production system

Billson Solar — Interactive AI Sales Engine

Interactive solar e-commerce system that calculates customer energy requirements and generates instant equipment recommendations and quotes.

Business Outcome:

Customers receive instant, accurate solar equipment recommendations tailored to their energy needs with live product specifications and pricing.

Architecture Stack:
Next.jsAI APIsTypeScriptTailwind CSS
Sales ProcessAutomated Sizing
Quote Conversion+3.5x Uplift
Explore Case Study
PROJECT 03 · AUTOMATION / KNOWLEDGE AIWorkflow and public UI evidence

Standout4Growth — Intelligent Brand Knowledge Assistant

Custom AI assistant that resolves customer onboarding and service inquiries 24/7 using verified company documentation with zero hallucinations.

Business Outcome:

Reduces repetitive client support triage by 70% while maintaining strict brand voice and private data security.

Architecture Stack:
n8nSupabaseGeminiLangChain
Support Triage70% Automated
Response Speed< 2 Seconds
Explore Case Study
PROJECT 04 · CYBER-DEFENSE / NIDSCybersecurity defense system

AI Integrated Network Threat Detection Engine

Machine-learning Network Intrusion Detection System analyzing live network flows to detect attacks and explain security risks in real time.

Business Outcome:

Evaluated on 82,332 held-out network flows with binary and multi-class threat classification, live PCAP parsing, and SOC-style alerts.

Architecture Stack:
PythonXGBoostStreamlitNFStreamScapy
Flows Evaluated82,332
Threat DetectionReal-Time
Explore Case Study
Verified Client Outcomes

What Operating Leaders Say.

Real feedback from business directors in Nairobi running mission-critical web platforms, automated AI pipelines, and zero-trust systems.

5.0 / 5.0
< 0.8s Mobile TTFB · 100% Uptime

“When we launched our citywide billboard campaigns across Nairobi, our biggest fear was server failure under sudden surge traffic. HarLyn engineered a Next.js platform that loaded in under 0.8s on mobile, with automated M-Pesa Daraja payment ledgering. We processed thousands of instalment applications with 100% uptime throughout the entire campaign.”

Kelvin M.
Head of Digital Operations · StarCall Phones Nairobi
Nairobi, Kenya · September 2026
View Evidence
5.0 / 5.0
+3.5x Quote Conversion · 30s Turnaround

“Before HarLyn built our interactive solar sizing engine, our sales engineers spent up to 48 hours manually calculating load profiles and generating customer quotes. Harrison and Nazline automated the entire discovery funnel. Clients now get tailored capacity recommendations and quotes in 30 seconds, driving a 3.5x boost in quote conversions.”

Brian O.
Commercial Operations Director · Bills On Solar East Africa
Nairobi / East Africa · August 2026
View Evidence
5.0 / 5.0
70% Inquiries Automated · Zero Hallucinations

“Our senior team was losing 20+ hours every week fielding repetitive onboarding and statutory compliance questions. HarLyn deployed a private knowledge assistant trained on our verified methodology. It resolves 70% of inbound triage instantly with zero hallucinations and strict data privacy under KDPA.”

Faith N.
Managing Partner · Standout4Growth Consulting
Nairobi, Kenya · August 2026
View Evidence
5.0 / 5.0
4 Bottlenecks Remediated · 48-Hr SLA

“We brought in HarLyn for their 48-Hour Digital Workflow Assessment. Within 48 hours, they uncovered 4 hidden bottleneck failure points in our dispatch routing and provided code-level fixes. Their direct founder model means you deal directly with elite engineers, not non-technical account managers.”

David K.
Operations & Logistics Lead · Munar Kenya
Nairobi / Mombasa · July 2026
View Evidence
100% Founder-Led Delivery· Harrison Ndeke (AI Systems) & Nazline Mwita (Cybersecurity)
Book 48-Hour Diagnostic (KSh 35,000 / $280)
Ecosystem Evidence

Proof across the whole partnership

HarLyn's missions above are one part of the record. Harrison and Nazline each maintain their own documented case studies on their personal sites — organized here by discipline, linking directly to the source.