Heftli · Swiss SaaS invoicing platform
Solo build, Jan 2026–Present
Problem Freelancers and SMEs in Switzerland needed a clean, compliant invoicing tool with native QR-Bill generation and time tracking without the overhead of enterprise billing platforms.
Owned Full product: architecture, infrastructure (Supabase, Stripe), backend, frontend, multi-tenant data model, payments integration, and compliance. Every layer built and operated independently.
Outcome Production-grade multi-tenant SaaS platform live and in use. End-to-end product engineering at professional standard, solo.
Visit Heftli →
Expats in Switzerland · relocation platform
Solo build, Jan 2025–Present
Problem Professionals relocating to Switzerland needed a reliable, practical resource, and the product needed to operate at near-zero cost while still handling dynamic features and lead capture.
Owned Architecture, development, infrastructure, content, and operations. Evolved from a static site to a full platform with background workers and dynamic features, using cloud primitives cost-efficiently.
Outcome Live, growing platform serving the international community in Switzerland. Proof of resourceful, cost-efficient product engineering.
Visit Expats in Switzerland →
Internal developer platforms
Enterprise cloud programmes, 2017–2024
Problem Engineers spent significant time each sprint on manual infrastructure setup. Developer lab provisioning took days.
Owned Built self-service infrastructure platforms and on-demand developer lab provisioning across AWS, Azure, and GCP. Designed, built, and iterated on both platforms.
Outcome Eliminated manual infrastructure setup overhead per sprint. Lab provisioning reduced from days to under 10 minutes across all three major clouds.
Cloud migration and platform standardisation
Multi-environment cloud migration
Problem Application teams needed a repeatable migration path without carrying manual provisioning, unclear ownership, or unmanaged cost growth into the target cloud.
Owned Migration architecture, landing-zone decisions, IaC patterns, environment baselines, and hand-off documentation.
Outcome Moved workloads onto a more repeatable platform with clearer ownership, safer deployment paths, and better cost visibility.
Related notes
Plan cloud migration without operational risk
Cloud cost optimisation programme
Multi-account cloud estate
Problem Cloud spend was growing faster than workload growth, with underused capacity, idle resources, and limited visibility into cost drivers.
Owned Spend analysis, right-sizing recommendations, reserved and spot capacity decisions, and ongoing cost review workflows.
Outcome Reduced waste and improved cost predictability without compromising production reliability.
Related notes
Practical cloud cost optimisation
Scaling and load-balancing architecture
High-traffic web platforms
Problem Traffic spikes and sustained editorial load required horizontal scaling and load distribution without single points of failure.
Owned Capacity planning, autoscaling configuration, load balancer design, and caching layers for web workloads.
Outcome Platforms held under sustained and peak traffic with clearer scaling paths and operational runbooks.
Golden image and hardening pipeline
Repeatable server provisioning
Problem Manual server builds introduced drift, inconsistent security baselines, and slow environment spin-up.
Owned Image build pipelines with Packer, baseline hardening controls, and versioned golden images for production environments.
Outcome Gave teams repeatable, hardened server baselines with faster provisioning and fewer configuration surprises.
RBAC and access governance
Production cloud environments
Problem Broad or ad hoc access made production changes harder to audit and increased the risk of privilege creep.
Owned Role design, least-privilege access patterns, access review workflows, and change-control guardrails.
Outcome Tightened production access with clearer ownership, auditable permissions, and safer operational hand-offs.
Serverless workload migration
Event-driven and API workloads
Problem Some services carried operational overhead that did not justify always-on infrastructure.
Owned Serverless architecture decisions, deployment patterns, and migration of suitable workloads off fixed capacity.
Outcome Reduced operational toil for event-driven workloads while keeping observability and recovery paths clear.
Batch processing pipeline
Operational and data workloads
Problem Scheduled and batch jobs needed reliable execution, retry behaviour, and clear failure handling without manual intervention.
Owned Batch job design, scheduling, failure recovery, and integration with existing data and messaging systems.
Outcome Delivered dependable batch processing with predictable run times and operational visibility.
Data lake and analytics platform
Large-scale data movement and discovery
Problem Operational and business data needed a central store for analysis without breaking existing production pipelines.
Owned Data lake architecture, ingestion patterns, storage design, and integration with search and batch workloads.
Outcome Enabled analytics and discovery over consolidated data while keeping production systems decoupled.
CI/CD standardisation programme
Multi-team application delivery
Problem Inconsistent build and deploy paths created drift, slow releases, and hard-to-debug production changes.
Owned Pipeline templates, deployment standards, environment promotion workflows, and team hand-off documentation.
Outcome Gave application teams repeatable CI/CD patterns with fewer manual steps and clearer release ownership.
Delivery quality with Playwright
Web application release confidence
Problem Critical user flows needed automated verification before production releases without slowing delivery.
Owned End-to-end test design, Playwright integration into CI pipelines, and failure triage workflows.
Outcome Increased release confidence for key user journeys with automated checks in the delivery path.
Bind9 DNS management interface
Production DNS operations
Problem Manual zone edits and ad hoc shell access made DNS changes risky and hard to audit.
Owned Custom zone management tooling, change workflows, and operator guardrails for Bind9.
Outcome Gave operators a controlled, auditable path for production zone updates without manual file edits.
API platform
Mobile apps and embedded hardware clients
Problem Multiple client types needed a stable integration surface that could evolve without breaking consumers.
Owned API design, backend services, and integration contracts for mobile and Raspberry Pi clients.
Outcome Delivered a versioned API boundary used across client types, with a public prototype on GitHub.
View prototype →
Elasticsearch search application
Large indexed dataset, filter-heavy discovery
Problem Users needed fast filtered discovery over a high-volume index with date and location aggregations.
Owned Search service design, aggregation queries, and Silex application layer over Elasticsearch.
Outcome Enabled sub-second filtered search with date and location aggregations over a high-volume index.
Multilingual news platform
High-traffic editorial publishing, multi-region readership
Problem Multi-language publishing had to stay fast and stable under sustained editorial traffic.
Owned Hosting architecture, localization pipeline, caching, and platform ops across Joomla and WordPress.
Outcome Kept multi-language publishing fast and stable under sustained editorial traffic.
Related notes
How to fix slow wordpress admin,
Secured wordpress on ubuntu 18.04 with nginx
Large crowdlearning platform
Community content at scale
Problem High-volume community publishing needed operational patterns that held under load.
Owned WordPress platform architecture, publishing workflows, caching, and production hardening.
Outcome Supported high-content-volume publishing with operational patterns that held under load.
Related notes
Hardening wordpress on lamp,
How to fix slow wordpress admin
Industrial CMS platform
Multi-property news publishing
Problem Multiple editorial properties needed a shared publishing backbone with repeatable deployment patterns.
Owned Shared platform design across PHP and Python components for editorial teams.
Outcome Gave multiple properties a common publishing backbone with repeatable deployment patterns.
Video search engine
Indexed media discovery product
Problem Media retrieval needed a modern front-end backed by a versioned API surface over an evolving stack.
Owned System architecture across CodeIgniter, Laravel, and Vue.js with an evolving API layer.
Outcome Shipped searchable media retrieval with a modern front-end backed by a versioned API surface.
Budget management platform
International logistics company
Problem Finance and operations teams needed a maintainable platform for budget tracking and sign-off.
Owned End-to-end delivery of reporting and approval workflows on Laravel.
Outcome Gave finance and operations teams a maintainable platform for budget tracking and sign-off.
Hackathon application
Competition delivery under time pressure
Problem A working product had to ship on deadline within hackathon constraints.
Owned Full-stack build on Laravel 4 within hackathon constraints.
Outcome Shipped a working product on deadline and received an award for the delivery.