Languages we work in, grouped by the work they&#x27;re best for.

Default choice for production web apps, full-stack Next.js, React Server Components, Node / Bun / Deno, and any typed frontend system. If it compiles, it runs.

Sample work

·Next.js customer portal with Postgres, auth, Stripe billing, and AI-assisted support workflows
·Type-safe SDK + OpenAPI integration layer for an existing Python backend

When to pick

Long-lived codebase, large team, refactoring-heavy domain.

When not to pick

Throwaway 50-line script or CPU-bound numerical work.

See details

JavaScript

Strong on existing browser, Node, and React codebases that haven’t migrated to TypeScript. We won’t force a rewrite to charge for one.

Sample work

·Modernize a legacy React / Node app without forcing a full TS migration
·Serverless function workflow on Vercel / Cloudflare Workers

When to pick

Existing JS codebase or serverless edge runtimes.

When not to pick

Greenfield project with a team that can adopt TypeScript instead.

Backend & API

5 languages

Python

Primary language for AI apps, RAG pipelines, eval harnesses, data workflows, and FastAPI / Django services. The lingua franca of the AI stack.

Sample work

·RAG ingestion pipeline with FastAPI, pgvector, eval datasets, and model routing
·Airflow + dbt pipeline that prepares LLM training data nightly

When to pick

AI / ML, data engineering, rapid API prototyping, scientific computing.

When not to pick

Sub-millisecond latency budgets or memory-safety-critical systems.

See details

Ruby

Strong on Rails codebases, marketplace apps, and admin systems. We don’t lead with Ruby on greenfield in 2026, but we ship clean Rails work on existing estates.

Sample work

·Add AI-assisted document processing to an existing Rails operations app
·Modernize a Rails admin panel with a React frontend + retain Rails as API

When to pick

Existing Rails monolith with a team already fluent.

When not to pick

Greenfield project or high-concurrency / ML-heavy workloads.

Java

We accept Spring Boot work, JVM enterprise integrations, and Android backend services. We will usually recommend Kotlin or Go on greenfield.

Sample work

·Spring Boot integration layer between an internal system and an LLM extraction service
·Java microservice that exposes an existing enterprise database via REST + observability

When to pick

Existing JVM estate or enterprise integration requirements.

When not to pick

Greenfield project where Go or Kotlin would deliver faster with less boilerplate.

Go

Built for concurrency. Our pick for high-throughput APIs, workers, CLIs, and services with single-binary deployment.

Sample work

·Data ingestion worker that syncs SaaS APIs into Postgres and a vector store
·gRPC service mesh + Kubernetes operator for a multi-tenant SaaS

When to pick

High I/O concurrency, single-binary deploys, simple-to-reason network services.

When not to pick

Complex inheritance domains or heavy numerical computation (use C++ / Rust).

PHP

Supportable for Laravel and existing PHP estates. Not a preferred greenfield stack unless the client already runs it.

Sample work

·Stabilize a Laravel business app and add structured data-export automation
·Modernize a WordPress + WooCommerce stack with a Next.js headless frontend

When to pick

Existing PHP estate or team fluency.

When not to pick

Greenfield project with no existing PHP investment.

Mobile

3 languages

Swift

Native iOS / macOS / iPadOS when the product requires Apple-platform depth — CoreML, ARKit, on-device inference, Catalyst.

Sample work

·iOS field-capture app with offline sync and on-device CoreML inference
·macOS menu-bar utility integrating with native AppKit + CoreData

When to pick

iOS-first product or best-in-class iOS UX expected.

When not to pick

Single-codebase cross-platform required or web-first product.

Kotlin

Native Android with Jetpack Compose. Also a strong JVM backend language with coroutines if you’re escaping Java boilerplate.

Sample work

·Android companion app with NNAPI on-device inference + offline sync
·JVM backend replacing legacy Java with clean Kotlin + coroutines

When to pick

Android-first product or JVM team that wants less Java boilerplate.

When not to pick

iOS-only product or tight memory-footprint constraint.

Dart

Our default cross-platform mobile pick via Flutter when one codebase must deploy native to iOS and Android.

Sample work

·Flutter MVP with auth, offline capture, and API sync — single codebase for iOS + Android
·Branded customer-facing app with pixel-perfect UI across both platforms

When to pick

Cross-platform mobile MVP with budget for one codebase, not two.

When not to pick

Deep platform-native features or heavy 3D / graphics workloads.

Data & Query

1 language

SQL

Core language for every serious data-centric app. We write raw Postgres SQL — window functions, RLS policies, JSONB — where ORMs choke.

Sample work

·Redesign a Postgres schema and dashboard queries for an internal reporting app
·Multi-tenant RLS policies + optimized window functions for SaaS analytics

When to pick

Reporting / analytics, multi-tenant data isolation, transactional integrity.

When not to pick

Unstructured blob storage or key-value-only workloads.

Systems & Performance

4 languages

C

Focused C work where direct hardware access, protocol-level work, or performance-critical paths matter. We don’t advertise safety-critical firmware (ISO 26262 / DO-178C).

Sample work

·Write or debug a C utility that parses device telemetry from an edge system
·Native C extension for a Python data pipeline where the GIL is a bottleneck

When to pick

Direct hardware manipulation or protocol parsers with small focused scope.

When not to pick

Web APIs, rapid prototyping, or anything where memory safety should be enforced by the compiler.

C++

Modern C++ (17 / 20) for performance-sensitive modules, embedded-adjacent tools, native SDKs, CMake / Conan build systems, and AI-inference engines (llama.cpp / ggml).

Sample work

·Port a Python ML inference path to llama.cpp / ggml C++ binaries for ~60% cost + latency reduction
·Profile and refactor a C++ processing module; clean up CMake; expose a Python binding

When to pick

Deterministic latency, legacy C++ estate, native SDK integration, performance-sensitive AI inference.

When not to pick

Web backend, MVPs that need PMF validation in 3 weeks, or safety-critical firmware.

See details

Rust

Memory-safe systems work without GC. Our pick for systems daemons, CLI tools, WebAssembly modules, and high-performance services where security and speed intersect.

Sample work

·Rust CLI that validates and transforms large structured files safely + concurrently
·WebAssembly module compiled from Rust for browser-side cryptography

When to pick

Memory-safety + C++ performance both required, greenfield systems, or WASM target.

When not to pick

MVP in 1 week or a team without bandwidth for the borrow checker.

Zig

Cutting-edge C drop-in replacement. We take focused Zig work for hyper-optimized memory layouts, cross-compilation, and C-interop without C’s undefined-behavior baggage.

Sample work

·Custom allocator + benchmark for a small native utility where C interop matters
·Cross-compile a parser to multiple target triples from a single Zig codebase

When to pick

Predictable binaries, C-interop without preprocessor pain, bleeding-edge stacks.

When not to pick

You need a mature package ecosystem or standard web / backend work.

Scripting & Glue

1 language

Bash

Practical for CI/CD pipelines, local tooling, deployment scripts, and developer-workflow automation.

Sample work

·Clean up build scripts and deployment automation for a multi-service app
·GitHub Actions workflows wired into Terraform deploys

When to pick

CI/CD glue, one-off automation, deploy scripts.

When not to pick

Complex application logic — escalate to Python or Go.