Open to Full-Stack & AI Engineering roles

Building production-grade enterprise software
and AI systems.

Enterprise software·Automation·AI systems·Built for production
Current focus:
Enterprise Platforms/Agentic AI/Distributed Systems
Start a conversation →View architecture ↓

Currently building

  • Agent Operations Platform
  • Enterprise CRM
  • AI Automation

Engineering Decisions

I build software that remains maintainable after the first release. I focus on system design, operational reliability, and business outcomes—not just shipping features.

Shipped Products

ProductionEnterpriseRealtime
2026Enterprise Logistics Platform

AGENT OPS PLATFORM

Role: Lead Full-Stack EngineerStatus: ProductionTeam: Solo
Enterprise logistics platform for fleet management, inventory orchestration, and AI-assisted operations.
Real-time geospatial tracking, RBAC
Vue 3·NestJS·Redis·PostgreSQL·Socket.IO·Prisma·Mapbox GL
Architecture →
ProductionEnterprise
2026Enterprise Platform

TINIQ CRM

Role: Lead Full-Stack EngineerStatus: ProductionTeam: Solo
Centralized sales operations hub
Eliminating lost context and manual data entry in sales pipelines.
Vue 3·Fastify·Redis·PostgreSQL·Socket.IO
Architecture →
Production
2024Corporate Website

MY DREAM

Role: Frontend EngineerStatus: ProductionTeam: Solo
High-speed, SEO-optimized corporate platform
Delivering dynamic-feeling content with zero backend infrastructure.
Astro·Svelte·TypeScript
Architecture →
ProductionFrontend
2025Frontend Engineering

FARID ZIYO

Role: Frontend EngineerStatus: ProductionTeam: Solo
High-performance brand digital presence
Fluid motion sequences with optimal Core Web Vitals.
Vue 3·CSS·Cloudflare
Architecture →

Engineering Case Studies

Tiniq CRM

+

Problem

Fragmented comms and manual entry caused lost context. Reps constantly refreshed dashboards to see updates.

Architecture

Vue 3 frontend via REST for CRUD and Socket.IO for real-time events. Fastify backend with Redis for caching/PubSub and PostgreSQL for persistence.

Constraints

  • Must support concurrent multi-user edits
  • High volume of real-time events
  • Strict role-based access control

Trade-offs

Redis Pub/Sub + Socket.IO over polling increased complexity but reduced API load by 80%.

Result

Centralized hub with zero manual refreshes. Reduced latency and isolated heavy tasks.

Lessons Learned

  • Real-time defaults force early focus on idempotency.
  • BullMQ task isolation saved the Node.js event loop.
  • Fastify schema validation dramatically reduced runtime errors.

Agent Ops Platform

+

Problem

Fleet management and inventory orchestration were disconnected. Geospatial tracking updates were delayed and unreliable.

Architecture

Vue 3 + Mapbox GL on the frontend. NestJS orchestration layer managing Socket.IO connections. Redis for geospatial caching and PostgreSQL/Prisma for persistent state.

Constraints

  • Real-time map rendering at 60fps
  • Handle massive geospatial data ingestion
  • Offline-capable driver interfaces

Trade-offs

Processing geospatial aggregations in Redis before PostgreSQL increased architecture complexity but ensured sub-100ms UI updates.

Result

Unified logistics dashboard with live fleet tracking, automated AI-assisted routing, and zero data loss during offline driver syncs.

Lessons Learned

  • Geospatial data requires dedicated caching strategies, not just generic KV stores.
  • NestJS strict typing across the full stack prevents costly payload mismatches.
  • Mapbox GL performance requires careful management of Vue reactivity.

MY DREAM

+

Problem

The client needed a blazing fast, easily maintainable site without the overhead of a CMS or database, but still required complex structured data presentation.

Architecture

Astro for static site generation. Svelte islands used strictly where interactivity is needed. All "database" content is structured as typed TypeScript arrays at build time.

Constraints

  • Zero backend infrastructure
  • Extremely fast page loads (static HTML)
  • Data must be easily updateable without database migrations

Trade-offs

Hardcoding data into arrays means developers must commit code to update content, but completely eliminates server costs and database latency.

Result

A lightning-fast website with 100/100 Lighthouse scores. Content is strongly typed and statically rendered, eliminating layout shifts.

Lessons Learned

  • Not every project needs a database. Static arrays with TypeScript interfaces often provide the best developer experience for read-heavy sites.
  • Svelte pairs perfectly with Astro due to its zero-runtime philosophy.
  • Build-time data validation is safer and faster than runtime API validation.

Farid Ziyo

+

Problem

A high-end personal brand required a digital presence that matched its premium status, demanding complex motion without sacrificing performance.

Architecture

Astro for static HTML generation and zero-JS routing. Vue 3 islands for interactive components. GSAP for hardware-accelerated scroll motion.

Constraints

  • Strict performance budget (LCP < 1.5s)
  • Complex, timeline-driven scroll animations
  • Perfect accessibility and SEO scores

Trade-offs

Separating animation logic from component reactivity (Vue + GSAP) required manual cleanup but completely eliminated dropped frames.

Result

A stunning, award-winning digital experience that scores 100/100 on Lighthouse while delivering complex 60fps motion.

Lessons Learned

  • Astro islands architecture forces you to be highly intentional about shipping JavaScript.
  • GSAP ScrollTriggers must be carefully unmounted to prevent memory leaks in SPAs.
  • Design tokens combined with CSS variables enable rapid, consistent UI iteration.

Engineering Impact

Realtime WebSocket layer
Realtime synchronization
Live operational visibility
Reduced manual work
Centralized workflow
[Agent Ops + Farid Ziyo GEC]
BullMQ queue decoupling
Reduced API latency
Responsive user experience
Background task isolation
[Tiniq CRM + Agent Ops]

Systems Architecture

A real production stack — not a demo. Each layer exists for a specific reason. Hover any node to see the engineering rationale and trade-off analysis.

Client

Browser

Vue 3 / GSAP
Why

Vue 3 provides reactivity; GSAP handles animation outside the component tree for performance.

Trade-off vs React

Benefits: Massive ecosystem and JSX flexibility

Limitations: Hooks can cause unpredictable renders in heavy real-time apps

Decision: Predictable reactivity for real-time data binding.

AI Engineering

I integrate language models into production workflows — as reasoning engines inside automated pipelines, not as chat wrappers.

Reasoning & Planning
State-aware agent planning.
Structured Outputs
Deterministic JSON generation.
Tool Calling
Secure backend tool execution.
Multi-Agent Workflows
Distributed agent orchestration.
AI Orchestration
Bridging AI and business logic.

Engineering Principles

01

Ship before polish

Production feedback beats local perfection.

02

Business before abstraction

Architecture exists to support product goals.

03

Measure before optimizing

Performance decisions require evidence.