Hire developers for your exact tech stack

React is how most product companies build their web frontend, from early-stage startups to Fortune 500 engineering orgs. The challenge is not finding someone who knows React. It is finding someone who knows when not to use useEffect, how to structure a component library that will not become a maintenance nightmare, and how to squeeze performance out of a heavy dashboard. NeuronHire places LATAM React engineers vetted on React 18+, TypeScript, Next.js, and real-world state management patterns. First profiles in 7 days, 30–50% below US rates.

Angular is the frontend framework for teams that need strong conventions enforced by the framework, not by code review. Its opinionated structure, strict TypeScript, and built-in dependency injection make it well-suited for large codebases where inconsistency compounds over time. Good Angular engineers who understand RxJS deeply, manage state without over-engineering, and know the post-v16 signals model are harder to find than Angular generalists. NeuronHire places Angular developers from Latin America vetted on Angular 17+, RxJS, NgRx, and enterprise patterns. Timezone-aligned with US teams and 30–50% below US rates.

Vue.js is the frontend framework product teams reach for when they want React's power without React's ceremony. The Composition API in Vue 3 is genuinely good, Pinia replaced Vuex cleanly, and Nuxt 3 is a serious full-stack framework. Where teams get into trouble is hiring Vue developers who know Vue 2 Options API and calling it done. The Vue 3 Composition API is a meaningfully different mental model and production experience. NeuronHire's LATAM Vue engineers are vetted on Vue 3 specifically: Composition API, Pinia, TypeScript integration, and Nuxt 3 in production. First profiles in 7 days, 30–50% below US rates.

Next.js is the React framework companies move to when they need SSR, fast page loads, and a real backend in a single codebase. The App Router and React Server Components changed how the framework works. Most developers who list Next.js on a resume have not shipped production apps using them. NeuronHire places Next.js developers from LATAM tested on App Router architecture, RSC data patterns, streaming, and full-stack Next.js. Not people who have only run create-next-app. Timezone-aligned with US teams, placed in 7 days, at 30–50% below US rates.

TypeScript is now the default for any JavaScript codebase with more than two developers. The static type system catches an entire class of runtime bugs before they reach production, and the IDE tooling alone makes teams faster. TypeScript depth varies widely. Plenty of engineers add basic type annotations and call it TypeScript. Real specialists design type-safe API boundaries with Zod, write conditional types that eliminate unsafe casting, and know when strict mode hurts more than it helps. NeuronHire's LATAM TypeScript engineers are vetted on generics, utility types, and applied TypeScript across React, Node.js, and full-stack systems. First profiles in 7 days, 30–50% below US rates.

JavaScript is the only language that runs natively in every browser and, through Node.js, on your backend too. The supply of JavaScript developers is large, but engineers who deeply understand the event loop, prototype chain, and async model are far rarer than the market suggests. NeuronHire places pre-vetted JavaScript developers from Latin America, assessed on modern ES2024+, async patterns, the event loop, and real application architecture, in 7 days at 30–50% below US rates.

Node.js is the dominant backend runtime for JavaScript-first teams. The event loop handles high-concurrency I/O workloads efficiently, startup time is fast enough for serverless, and sharing types across the full stack removes an entire category of bugs. The gap between a developer who knows Node.js and one who understands the event loop, async error propagation, and memory management is significant. NeuronHire places Node.js developers from LATAM vetted on Express, NestJS, production API design, and async patterns under real load. Placed in 7 days, at 30–50% below US rates.

Python is the dominant language across three disciplines: backend APIs, data engineering, and ML/AI. A strong FastAPI engineer and a strong PySpark engineer both write Python, but they are not interchangeable. NeuronHire places Python developers from LATAM vetted by discipline: backend (Django, FastAPI), data engineering (Pandas, PySpark, Airflow), and ML/AI (PyTorch, LangChain, scikit-learn). Engineers are timezone-aligned with US teams, placed in 7 days, and cost 30–50% less than US equivalents.

Java dominates enterprise backend engineering. Banking, insurance, e-commerce, and large-scale SaaS platforms run on it because the Spring ecosystem, JVM performance characteristics, and decades of production hardening make it the default when reliability and team size both matter. NeuronHire places pre-vetted Java developers from Latin America, assessed on Spring Boot, microservices patterns, JPA/Hibernate, and cloud-native Java, in 7 days at 30–50% below US rates.

.NET powers the APIs, identity systems, and Azure-integrated services that keep enterprise SaaS and fintech platforms running. The US market forces companies to wait months for a senior .NET hire at a price that strains engineering budgets. NeuronHire places pre-vetted .NET engineers from Latin America, aligned to your timezone, fluent in C# 12+, ASP.NET Core, EF Core, and modern .NET 8+ patterns. Placed in 7 days at 30–50% below US rates.

Go is a statically typed language built for high-concurrency infrastructure: fast startup, small binaries, and predictable performance under load. Kubernetes, Docker, and Terraform are all written in Go. NeuronHire places pre-vetted Go developers from Latin America, assessed on idiomatic Go, goroutines and channels, gRPC, and cloud-native patterns, in 7 days at 30–50% below US rates.

PHP runs more than 75% of the web, including WordPress, WooCommerce, Drupal, and a large share of SaaS businesses built on Laravel. Engineers who dismiss PHP have not looked at it since 2012. Modern PHP 8.3 is typed, fast, and nothing like the PHP that earned its bad reputation. NeuronHire places PHP developers from LATAM vetted on PHP 8.3+, Laravel, Symfony, and production API design, including developers who specialize in modernizing the legacy codebases most companies actually have. Placed in 7 days, at 30–50% below US rates.

Rails is still the fastest way to build a full-featured web product from scratch. Convention over configuration means a small team ships in weeks what takes months in other stacks. GitHub, Shopify, and Basecamp still run on it. Rails 7 with Hotwire has made the modern frontend question largely irrelevant for most SaaS products. The shortage is not Rails developers. It is Rails developers who know the performance traps, the N+1 landmines, and how to keep a five-year-old codebase from becoming a liability. NeuronHire's LATAM engineers are vetted on exactly that. First profiles in 7 days, 30–50% below US rates.

Swift is Apple's programming language for iOS, macOS, watchOS, and tvOS. If your product lives on iPhone, there is no shortcut: Swift developers need deep platform knowledge of SwiftUI, Combine, CoreML, and the App Store submission process. NeuronHire places pre-vetted iOS developers from Latin America, assessed on Swift 5.9+, SwiftUI, UIKit, Combine, and App Store deployment, in 7 days at 30–50% below US rates.

Native Android development with Kotlin is the right choice when your app depends on hardware integration, Google services, or performance that cross-platform frameworks cannot match. Kotlin is Google's official Android language: null-safe, concise, and built around first-class coroutine support. NeuronHire vets Android Kotlin developers from Latin America on MVVM/MVI architecture, Jetpack Compose, Coroutines, and Play Store release workflows. They are timezone-aligned with US teams and cost 30–50% less than equivalent US hires.

React Native lets one team ship to both iOS and Android without sacrificing native feel, which is why it is the default mobile framework for most product teams that are not building something deeply platform-specific. Maintaining separate iOS and Android teams is expensive. The gap in React Native experience is real: plenty of developers can build basic screens, but far fewer know how to debug bridge performance, write a native module, or navigate EAS Build configuration. NeuronHire's LATAM engineers are vetted on exactly those areas. First profiles in 7 days, 30–50% below US rates.

Flutter lets you ship pixel-perfect, high-performance apps across mobile, web, and desktop from a single Dart codebase without the UX compromises of earlier cross-platform tools. Maintaining separate iOS and Android codebases doubles your mobile engineering cost. For most product teams, that trade-off is not worth making. NeuronHire places pre-vetted Flutter developers from Latin America vetted on Flutter 3+, Riverpod, BLoC, and multi-platform deployment. Placed in 7 days at 30–50% below US rates.

Apache Spark is the distributed computing engine serious data engineering and ML teams reach for when data volume exceeds what single-node tools can handle: terabytes of raw events, petabyte-scale joins, ML training on datasets that do not fit on one machine. The gap between a Spark engineer who can write a PySpark job and one who can tune it is enormous. Partition management, shuffle avoidance, and join strategy optimization show up directly in cluster costs and pipeline SLAs. NeuronHire's LATAM Spark engineers are vetted on PySpark, Structured Streaming, Delta Lake, and Databricks deployment. First profiles in 7 days, 30–50% below US rates.

Apache Airflow is the leading open-source workflow orchestration tool for scheduling, running, and monitoring Python-defined data pipelines. Companies hire Airflow engineers when cron jobs break silently and nobody knows why. Most teams hit Airflow's complexity wall fast: custom operators, dynamic DAGs, Celery vs. Kubernetes executors. NeuronHire places Airflow specialists from Latin America vetted on DAG design, Astronomer, MWAA, and Cloud Composer deployments. They are timezone-aligned with US teams and cost 30–50% less than equivalent US hires.

Snowflake is the cloud data warehouse most analytics-forward companies land on when they outgrow Postgres or BigQuery. The elastic compute model means you are not paying for idle capacity. The separation of storage and compute means your BI team is not fighting your ETL jobs for resources. Snowflake is easy to get running and expensive to get wrong. Warehouse sizing, clustering keys, and query patterns have a direct impact on your monthly bill. NeuronHire's LATAM Snowflake engineers are vetted on SQL optimization, data modeling, Snowpark, and dbt integration. First profiles in 7 days, 30–50% below US rates.

Databricks is the unified platform for organizations that need serious data processing volume and ML in one place. Spark, Delta Lake, MLflow, and Unity Catalog work together in a way that separate point solutions cannot match. Databricks expertise runs deep: Spark performance tuning, Delta Live Tables design, Unity Catalog governance, and cluster cost management are all distinct skills. NeuronHire places Databricks engineers from Latin America vetted on Spark, Delta Lake, Workflows, Unity Catalog, and MLflow. Placed in 7 days at 30–50% below US rates.

Apache Kafka is a distributed event streaming platform for real-time data pipelines, event-driven microservices, and event sourcing at scale. It handles millions of events per second with durable, ordered logs that multiple consumers can read independently. Kafka requires engineers who deeply understand partitioning, consumer group design, and exactly-once semantics, or subtle data loss bugs reach production. NeuronHire places pre-vetted Kafka engineers from LATAM in 7 days at 30–50% below US rates, assessed on Confluent Platform, Kafka Streams, Schema Registry, and cluster operations.

dbt is the standard for SQL-based transformations in the modern data stack because it brings software engineering practices: version control, testing, documentation, and CI/CD to analytics work that previously lived in personal SQL scripts. An experienced dbt engineer designs a project structure that scales, writes tests that catch real data quality issues, and keeps metric definitions consistent across every downstream consumer. NeuronHire places dbt engineers from Latin America vetted on dbt Core, dbt Cloud, dimensional modeling, Snowflake, BigQuery, and Redshift. Placed in 7 days at 30–50% below US rates.

Building a proof of concept with the OpenAI API takes an afternoon. Building a production AI feature that is reliable, cost-efficient, and does not hallucinate on your customers takes a real engineer. The work is in prompt architecture, output validation, RAG pipeline design, cost control, and observability. NeuronHire places OpenAI API specialists from LATAM vetted on the full stack: GPT-4o chat completions, the Assistants API, function calling, structured outputs, embeddings, and fine-tuning for domain-specific accuracy. First candidates in 7 days, at 30–50% below US rates.

Claude Code is Anthropic's agentic coding CLI and IDE integration that lets Claude autonomously read, write, run, and debug code inside your actual development environment. Engineers who specialize in Claude Code build production automation: MCP servers that connect Claude to internal tools, hooks that enforce workflow policies, and headless pipelines that run in CI without human intervention. NeuronHire places Claude Code specialists from Latin America vetted on agentic workflows, MCP server development, and CI/CD automation. Placed in 7 days at 30–50% below US rates.

Gemini is Google's LLM family with native multimodal understanding and a 2 million token context window. For teams on Google Cloud, Vertex AI makes it the default path for enterprise AI features. NeuronHire places pre-vetted Gemini engineers from Latin America, assessed on the Gemini API, Vertex AI pipelines, multimodal workflows, RAG, and grounding capabilities, in 7 days at 30–50% below US rates.

OpenClaw is the first autonomous agent runtime that works for real-world tasks: reading codebases, running tests, browsing the web, and calling APIs with an LLM as the decision-maker. It crossed 250,000 GitHub stars in under four months, the fastest adoption curve of any developer framework in recent memory. Engineers who know how to deploy it safely and extend it for enterprise use are rare. NeuronHire places OpenClaw engineers from LATAM vetted on tool integration, multi-agent coordination, and secure production deployment including NemoClaw sandboxing. First candidates in 7 days, at 30–50% below US rates.

CrewAI is an open-source Python framework for building multi-agent AI systems where each agent has a defined role, a set of tools, and a specific job in a larger workflow. The jump from a working demo to a production multi-agent system is significant. You need engineers who understand task delegation patterns, tool reliability, inter-agent communication, and how to handle the failure modes LLMs introduce at scale. NeuronHire places CrewAI engineers from Latin America vetted on crew design, tool integration, and production deployment. Placed in 7 days at 30–50% below US rates.

LangChain is a Python framework for building LLM applications: it provides abstractions for chains, agents, retrieval, and memory that cut the boilerplate of connecting LLMs to your data and tools. It is the most widely adopted LLM framework with the largest integration ecosystem. LangChain's abstractions can hide complexity that causes production failures in context management, retrieval quality, agent reliability, and cost control. NeuronHire places senior LangChain developers from LATAM in 7 days at 30–50% below US rates, vetted on LangGraph, LangSmith, RAG pipeline architecture, and multi-agent design.

LangGraph is the standard framework for building production agentic AI systems that need reliable state management, conditional branching, and human-in-the-loop control. Basic LangChain agents break on anything more complex than a few tool calls because they lack state persistence and reliable branching. LangGraph models agent workflows as graphs where nodes are execution steps and edges are conditional transitions. NeuronHire places pre-vetted LangGraph engineers from LATAM in 7 days at 30–50% below US rates, assessed on graph design, state management, multi-agent coordination, and LangGraph Cloud deployment.

LlamaIndex is the Python framework purpose-built for connecting LLMs to private data: contracts, documentation, financial records, and support tickets. It manages the full data pipeline from ingestion through chunking, embedding, indexing, retrieval, and synthesis. For pure RAG work, LlamaIndex offers more opinionated, production-ready abstractions than LangChain. Basic vector search rarely delivers production accuracy. Chunking strategy, reranking, and hybrid retrieval separate reliable systems from expensive demos. NeuronHire places pre-vetted LlamaIndex engineers from LATAM in 7 days at 30–50% below US rates, assessed on advanced retrieval, query engines, data agents, and RAGAS evaluation.

Pinecone is the managed vector database most teams use to store and retrieve embeddings in production AI applications. It handles the infrastructure, scales to billions of vectors, and does not require managing your own ANN index. The real engineering work is in the RAG architecture around it: chunking strategy, embedding model selection, namespace design, hybrid search, and metadata filtering. NeuronHire places vector database engineers from LATAM vetted on Pinecone, pgvector, and the full RAG stack. First candidates in 7 days, at 30–50% below US rates.

LangSmith is the observability, evaluation, and improvement platform for LangChain and LangGraph applications. Running LangChain in production without LangSmith means debugging blind: you cannot see which prompts fired, what the model returned, or where your pipeline is losing accuracy. LangSmith traces every LLM call with full context, lets you build evaluation datasets from real traffic, and catches prompt regressions before users notice. NeuronHire places pre-vetted LangSmith engineers from LATAM in 7 days at 30–50% below US rates, assessed on tracing integration, evaluator design, automated testing workflows, and prompt management.

LangFuse is an open-source LLM observability platform that provides production tracing, prompt management, evaluation workflows, and cost tracking for any LLM framework. Its self-hosted deployment option keeps every trace inside your own infrastructure, which is why teams in healthcare, finance, and government choose it over LangSmith. NeuronHire places pre-vetted LangFuse engineers from LATAM in 7 days at 30–50% below US rates, assessed on SDK integration, evaluation pipeline design, self-hosted deployment, and LLMOps workflows.

MLflow brings engineering discipline to the ML lifecycle through experiment tracking, model versioning, and a deployment-ready model registry. Data science teams without a disciplined tracking system spend more time re-running experiments than building models, and cannot reliably reproduce the run that actually worked. MLflow is also the standard in Databricks environments. NeuronHire places pre-vetted MLflow engineers from LATAM in 7 days at 30–50% below US rates, assessed on Tracking, Model Registry, Databricks managed MLflow, and LLM evaluation with MLflow 2.x.

Weights & Biases is the experiment tracking platform serious ML teams run, because training a model without it is like deploying code without logs. Every training run gets logged, every hyperparameter gets tracked, every artifact gets versioned. Teams that use W&B properly go further: Sweeps automate hyperparameter search, Artifacts version datasets and models, and W&B Weave traces and evaluates LLM applications. NeuronHire's LATAM engineers are vetted on the full W&B stack, including Weave for LLM tracing. First profiles in 7 days, 30–50% below US rates.

n8n is the automation platform technical teams reach for when they have outgrown Zapier but don't want to build everything from scratch. It is self-hostable, code-extensible, and has native LLM nodes for OpenAI and Anthropic. The right engineer wires together AI-powered business workflows in days, not months. NeuronHire places n8n engineers from LATAM vetted on workflow architecture, custom node development in TypeScript, AI and LLM integration, and self-hosted Docker deployments. First candidates in 7 days, at 30–50% below US rates.

Make is the automation platform teams reach for when they have outgrown Zapier's linear model and need routers, iterators, aggregators, and real error handling across multi-branch workflows. Its native AI modules for OpenAI and Anthropic make it a practical platform for teams shipping AI-enhanced workflows without a full engineering build. NeuronHire places pre-vetted Make engineers from LATAM in 7 days at 30–50% below US rates, vetted on advanced scenario design, AI module integration, error handling, and enterprise automation patterns.

Zapier connects 7,000+ apps and handles the automation work that would otherwise fall between engineering sprints or get done manually every week. The platform has grown well beyond simple two-step Zaps. AI Actions, Tables, Interfaces, and multi-path branching logic mean a skilled Zapier engineer can automate complex ops workflows without writing code. The difference between a basic Zapier user and a Zapier engineer shows up in error handling, logic branching, webhook architecture, and building automations that survive edge cases. NeuronHire's LATAM engineers are vetted on advanced Zap design, AI integration, and enterprise automation patterns. First profiles in 7 days, 30–50% below US rates.

TensorFlow is Google's production ML platform: the backbone of large-scale model training and serving at companies that need repeatable, auditable ML pipelines, not just notebooks that run once. While PyTorch dominates research, TensorFlow and TFX own a large share of enterprise ML infrastructure, especially on Google Cloud. The engineers who matter are not the ones who can write a training loop. They are the ones who can design a TFX pipeline, tune distributed training across TPUs, and ship a model to TensorFlow Serving without a 400ms latency regression. NeuronHire's LATAM engineers are vetted on exactly those skills. First profiles in 7 days, 30–50% below US rates.

PyTorch is the framework behind Llama, Mistral, Stable Diffusion, and most serious ML research that reaches production. If your team is training models, fine-tuning LLMs, or building production ML systems, PyTorch is almost certainly what you run. Finding engineers who have actually debugged distributed training jobs, optimized CUDA memory, and shipped models without blowing up latency is the hard part. NeuronHire's LATAM engineers are vetted on those depth areas: PEFT, LoRA, FSDP, torch.compile, and TorchServe. First profiles in 7 days, 30–50% below US market rates.

Hugging Face is the standard toolkit for teams building AI features on open-source models instead of paying per-token to OpenAI. The Transformers library and model hub give engineers access to Llama, Mistral, Gemma, Whisper, and thousands of other models, with PEFT fine-tuning, quantization, and production inference tooling built in. NeuronHire places pre-vetted Hugging Face developers from Latin America, assessed on the Transformers library, PEFT/LoRA fine-tuning, model deployment, and Inference Endpoints, in 7 days at 30–50% below US rates.

AWS engineers keep your infrastructure running and your cloud bill under control. The gap between an engineer who can click through the console and one who designs a secure, cost-efficient, multi-account architecture with IaC from day one is enormous. NeuronHire places AWS engineers from Latin America vetted on core services, Terraform, EKS, security best practices, and production architecture. They are timezone-aligned with US teams and cost 30–50% less than equivalent US hires.

Google Cloud Platform is the leading choice for data-intensive and AI-heavy workloads. BigQuery, Vertex AI, and GKE are best-in-class for their categories. NeuronHire places pre-vetted GCP engineers from Latin America, assessed on GKE, BigQuery, Dataflow, Vertex AI, and Terraform, in 7 days at 30–50% below US rates.

Azure is the logical cloud extension for companies already running on Microsoft infrastructure: Active Directory, Office 365, SQL Server, and .NET. Azure engineers who can navigate Entra ID, AKS, Azure DevOps pipelines, and IaC with Bicep or Terraform are the engineers who make enterprise cloud migrations actually land. NeuronHire places Azure engineers from Latin America vetted on AKS, Azure Functions, Azure DevOps, identity management, and Azure security best practices. They are timezone-aligned with US teams and cost 30–50% less than equivalent US hires.

Docker is the baseline for modern software delivery. Without containerization, you have environment drift, inconsistent CI builds, and a deployment process that depends on tribal knowledge. There is a wide gap between someone who can write a Dockerfile and someone who designs production-grade containers: multi-stage builds that minimize attack surface, optimized layer caching that cuts CI time significantly, and security hardening that passes a container audit. NeuronHire places Docker engineers from Latin America vetted on Dockerfile optimization, Docker Compose, container security, and CI/CD integration. Placed in 7 days at 30–50% below US rates.

Kubernetes is how serious engineering teams run containerized workloads in production: handling deployment, scaling, self-healing, and networking across cloud infrastructure without manual server intervention. A misconfigured RBAC policy or a missing resource limit can take down a cluster, and debugging a failing pod across namespaces requires genuine expertise. NeuronHire places senior Kubernetes engineers from LATAM in 7 days at 30–50% below US rates, vetted on Helm, GitOps with ArgoCD, EKS/GKE/AKS, RBAC, and platform engineering patterns.

Terraform is how serious engineering organizations manage cloud infrastructure: replacing manual console configurations with versioned, reviewable, repeatable code. The difference between a team that uses Terraform and one that uses it well shows up in incident response. When infrastructure is code, you can audit every change, roll back a bad deployment, and spin up a staging environment that exactly mirrors production. NeuronHire's LATAM Terraform engineers are vetted on module design, state management, Terragrunt, and multi-cloud patterns. First profiles in 7 days, 30–50% below US rates.

MySQL still runs the backends of some of the highest-traffic systems on the internet. That means the bar for who you hire to touch it is high. A bad index or a poorly designed replication setup does not surface in dev. It surfaces at 3am on Black Friday. NeuronHire places MySQL engineers from LATAM vetted on InnoDB internals, query optimization with EXPLAIN, replication topology, and managed MySQL on RDS and PlanetScale. First candidates arrive in 7 days, at 30–50% below US rates.

Redis appears in almost every high-performance production stack: as a cache, a session store, a rate limiter, a job queue, a pub/sub broker, or all five at once. The risk is not adopting Redis. It is adopting it without someone who understands cache invalidation pitfalls, cluster failover, and what happens to your data when persistence is misconfigured. NeuronHire's LATAM Redis engineers are vetted on data structures, cluster mode, Pub/Sub, Lua scripting, and Redis Stack. First profiles in 7 days, 30–50% below US rates.

PostgreSQL is the default database for most new production applications. ACID transactions, JSONB, full-text search, logical replication, and pgvector for AI workloads are all first-class. The gap between a developer who can write SQL and an engineer who designs schemas that perform at scale, tunes indexes correctly, and sets up replication without data loss is significant. NeuronHire places PostgreSQL engineers from LATAM vetted on schema design, query optimization with EXPLAIN ANALYZE, partitioning, replication, and pgvector. Placed in 7 days at 30–50% below US rates.

MongoDB is the right database when your data is naturally document-shaped, your schema is still evolving, or you need horizontal sharding that PostgreSQL cannot provide without significant engineering overhead. The risk is that MongoDB's flexibility cuts both ways. Teams without experienced engineers end up with schema designs that kill query performance or aggregation pipelines that take minutes instead of milliseconds. NeuronHire places pre-vetted MongoDB engineers from LATAM in 7 days at 30–50% below US rates, vetted on schema design, aggregation pipelines, Atlas, indexing strategies, and MongoDB Atlas Search.

Bad search kills conversion, and most teams only discover how badly their Elasticsearch cluster is misconfigured when load spikes or relevance complaints pile up. Elasticsearch powers full-text search, log analytics, APM, and vector similarity search across some of the highest-traffic systems in the world. NeuronHire places pre-vetted Elasticsearch engineers from Latin America who know index design, query DSL, ELK/Elastic Stack operations, and production tuning. Placed in 7 days at 30–50% below US rates.