Technology

The technology stack
behind what we ship

Proven platforms, custom-built tooling, and engineering practices refined across dozens of production deployments — from MVPs to enterprise scale.

Explore the stack Talk to an engineer
Backend
.NET 9
Mobile
Swift · Kotlin · Flutter
AI
RAG · Agents · MCP
request response Typical product, simplified
Core stack

Battle-tested foundations

We choose technologies for longevity, performance, and team velocity — not hype cycles. Pick an area to see what we use and why.

  1. 01Boring technology in the core, new ideas at the edges.
  2. 02One typed contract from database to screen.
  3. 03Every choice has to survive five years of maintenance.

Backend

The core of almost everything we ship: typed, fast and boring in the best way.

.NET 9 · ASP.NET Core

APIs, SaaS backends and admin panels — one language from domain model to deployment.

Primary

Entity Framework Core

Data access, migrations and query performance work on SQL Server and PostgreSQL.

Primary

SignalR & WebSockets

Live dashboards, chat, notifications and collaborative screens.

Real-time

JWT · OAuth 2.0 · OpenID

Sign-in, roles and API security, plus rate limiting and audit trails.

Security

Background workers

Hosted services, queues and scheduled jobs for anything that shouldn't block a request.

Async

Web

Interfaces that load fast, rank well and stay maintainable as the product grows.

React

Product UIs, dashboards and design systems built from reusable components.

Primary

Next.js

Server rendering and static generation for SEO-critical marketing and content.

Primary

TypeScript

Typed front-ends that share contracts with the .NET API.

Default

ASP.NET MVC · Razor

Server-rendered corporate sites and admin panels — this site included.

Content

Mobile

Native when the product needs depth, cross-platform when speed and shared UI win.

Swift · SwiftUI

iOS apps with full platform APIs, widgets and polished interaction.

Native

Kotlin · Jetpack Compose

Android apps tuned for performance across a wide device range.

Native

Flutter

One codebase, two stores — fast iteration for SaaS and education apps.

Cross-platform

React Native · Expo

Shared logic with React web products and over-the-air updates.

Cross-platform

Data

Relational at the core, with caching, search and analytics where they earn their place.

SQL Server

Transactional systems, reporting and long-lived enterprise data.

Primary

PostgreSQL

Open-source workloads, JSON documents and vector search with pgvector.

Primary

Redis

Caching, sessions, rate limits and lightweight queues.

Cache

Vector & hybrid search

Embeddings plus keyword search for grounded AI answers over your own data.

AI data

Analytics & observability

Product analytics, logs, traces and alerts wired in from day one.

Ops

Cloud & DevOps

Repeatable environments and releases that don't need a hero on call.

AWS

Compute, storage, CDN and managed databases for products of any size.

Cloud

Microsoft Azure

App Service, SQL and identity for Microsoft-centric organisations.

Cloud

Docker

The same container from a laptop to staging to production.

Default

Kubernetes

Orchestration for services that need horizontal scale and self-healing.

Scale

GitHub Actions

Build, test and deploy pipelines on every push — staging first, then production.

CI/CD

IIS · Windows Server

Battle-tested hosting for .NET workloads that live on-premises or in VMs.

Hosting

AI

LLM features engineered for production — measured, guarded and cost-aware.

OpenAI · Anthropic

Frontier models through their APIs, chosen per task on quality, latency and cost.

Models

Open-weight models

Self-hosted or small language models when privacy or cost calls for it.

Models

RAG pipelines

Chunking, embeddings, reranking and citations over your documents.

Retrieval

Agents & MCP

Tool-using agents and MCP servers with human approval and audit logs.

Agents

Evals & guardrails

Eval suites in CI, prompt-injection defences and PII redaction.

Quality
Mobile engineering

Native, cross-platform,
right tool for the job

We ship iOS and Android in production every week — not with a single framework religion, but with the stack that fits performance, timeline, and team constraints.

Swift & Kotlin when the product needs platform depth. Flutter & React Native when speed and shared UI win. Often both — in the same portfolio.

Native

iOS & Android

Maximum performance, platform APIs, and polish for apps where every frame and gesture matters.

Swift Kotlin SwiftUI Jetpack Compose
Cross-platform

Shared codebases

One team, two stores — rapid iteration for SaaS products, education platforms, and consumer apps at scale.

Flutter React Native .NET MAUI Expo
Ship & scale

Production mobile ops

Store releases, push notifications, offline sync, analytics, and crash monitoring — wired into the same CI/CD as backend.

App Store Connect Google Play Firebase Fastlane
Architecture

Built to scale from day one

Modular monoliths where they fit. Microservices when the domain demands it. We design for clear boundaries, testability, and zero-downtime deployments.

  • RESTful & real-time APIs with versioning and rate limiting
  • Event-driven patterns for async workflows and integrations
  • CDN-backed static assets and edge caching strategies
  • Automated backups, health checks, and incident playbooks
// gidysoft.engineering
const stack = {
  backend: "ASP.NET Core 9",
  mobile: {
    native: ["Swift", "Kotlin"],
    cross: ["Flutter", "React Native"]
  },
  web: "React · Next.js",
  cloud: "AWS · Azure",
  ai: "OpenAI · RAG · Agents"
};
AI engineering

AI that ships to production — not just to a demo

We design, build and operate LLM-powered products: agents that use your tools, search that understands your data, and the evals and guardrails that keep them honest.

  • ModelsOpenAI · Anthropic · open-weight
  • Stack.NET · vector search · MCP · cloud
  • ChannelsWeb, WhatsApp, Instagram & voice
  • QualityEvals in CI · guardrails · cost budgets
Build with AI AI services

AI agents & MCP

Multi-step agents with tool use, MCP servers, human-in-the-loop approval and full audit trails.

RAG & hybrid search

Vector + BM25 retrieval, filters and reranking — grounded answers over your own documents and data.

LLM product features

Copilots, summarisation, extraction and structured output wired into real workflows, not demos.

Voice & multimodal

Real-time voice agents, vision and speech pipelines tuned for latency, barge-in and honest fallbacks.

Evals & guardrails

Eval suites in CI, prompt-injection defences, PII redaction and cost budgets before anything ships.

Open & on-device models

Hosted APIs or open-weight and small language models — chosen on accuracy, latency, cost and privacy.

Ready to architect
your next product?

Let's talk stack, timeline, and what success looks like.

Start a conversation Explore services