We source, vet, and manage hiring so you can meet qualified candidates in days, not months. Strong English, U.S. time zone overlap, and compliant hiring built in.
.svg){kind=logo}
Jsonnet is a superset of JSON that adds variables, functions, conditionals, and imports. It compiles to plain JSON, making it ideal for generating large, repetitive configuration files without the boilerplate. Developed by Google and adopted by companies like Databricks, Grafana, and AWS, Jsonnet is the go-to tool for managing infrastructure-as-code complexity.
Unlike YAML-based approaches, Jsonnet treats configuration as code. You get type safety, reusability, and programmatic flexibility. A single Jsonnet template can generate hundreds of Kubernetes manifests, Terraform variables, or Grafana dashboards with minimal duplication. The language is Turing-complete, meaning you can express any logic needed to generate your configuration.
Jsonnet adoption is growing in the DevOps and SRE spaces. Major tools like ArgoCD, Ksonnet (the Kubernetes templating framework), and Grafana's jsonnet library have made it a standard skill for infrastructure teams. On GitHub, Jsonnet repositories show consistent growth in usage, particularly in the cloud-native ecosystem.
Stack position: Configuration language and templating tool for infrastructure, primarily used by DevOps engineers, SREs, and platform engineers building scalable configuration management systems.
Hire a Jsonnet specialist when you're building or scaling infrastructure-as-code systems, especially if you're managing multiple environments (dev, staging, production) or generating thousands of Kubernetes manifests. If your team is drowning in YAML duplication or manually managing configuration drift across clusters, Jsonnet eliminates that pain through programmatic generation.
Jsonnet is the right choice for teams running multi-tenant Kubernetes clusters where you need dynamic resource allocation based on customer tier, region, or workload type. It's particularly valuable in companies like SaaS providers, fintech platforms, and any organization managing complex microservices deployments where configuration consistency is critical.
Use Jsonnet when you need to generate not just Kubernetes manifests but also Terraform configurations, monitoring rules (Prometheus alerts), dashboard definitions (Grafana), or infrastructure CI/CD pipelines. It becomes essential when you have hundreds of similar resources that differ only in minor parameters (environment, region, customer).
Jsonnet is NOT a good fit if you're just managing simple, static configurations or if your team is unfamiliar with programming concepts. It adds cognitive overhead and requires developers comfortable with logic, recursion, and testing. Start with simple YAML templates first if your team isn't ready for a full templating language.
Team composition: Pair a Jsonnet specialist with Kubernetes engineers, DevOps practitioners, and infrastructure architects. They'll collaborate on schema design, reusable library patterns, and validation logic. In larger teams, Jsonnet becomes the foundation of a platform engineering layer.
Evaluate candidates on three dimensions: Jsonnet proficiency, infrastructure understanding, and software engineering practices. A strong Jsonnet developer doesn't just know the syntax; they understand template design patterns, DRY principles applied to infrastructure, and how to build maintainable configuration libraries.
Red flags: Candidates who treat Jsonnet as a simple text templating tool rather than a programming language, or who don't understand the difference between compilation-time and runtime evaluation. Also watch for developers who've never built reusable configuration libraries or who can't articulate their approach to testing generated configurations.
Strong signals: Portfolio projects showing well-organized Jsonnet libraries with clear abstractions, experience generating manifests for multiple Kubernetes clusters or cloud providers, and understanding of best practices like parameterization over duplication. Look for developers who've contributed to open-source configuration projects or maintained internal frameworks.
Junior (1-2 years): Should understand JSON and basic templating concepts, be able to write simple Jsonnet scripts that avoid duplication, understand conditionals and loops, and demonstrate familiarity with one infrastructure tool (Kubernetes, Terraform, or similar). Should be able to compile and debug Jsonnet locally and understand the difference between data and logic.
Mid-level (3-5 years): Should design reusable Jsonnet libraries and abstraction patterns, handle complex configuration generation with nested conditionals and recursion, write testable Jsonnet code, optimize for maintainability and performance, and have real-world experience managing multi-environment configurations. Should understand tradeoffs between parameterization and simplicity.
Senior (5+ years): Should architect large-scale configuration systems, design internal DSLs using Jsonnet, mentor teams on best practices, optimize Jsonnet code for performance (compilation time), and make strategic decisions about when to use Jsonnet versus other tools. Should have experience building configuration frameworks that thousands of developers rely on.
Soft skills matter: Remote infrastructure work requires strong async communication, documentation discipline, and the ability to explain configuration decisions to non-technical stakeholders. Jsonnet developers must communicate complex decisions clearly in code reviews.
Tell me about a time you eliminated configuration duplication in a project. What tool or approach did you use, and what was the impact? Good answers describe specific pain points, quantify the reduction in configuration files or lines of code, and explain how the solution scaled to other teams. Strong candidates mention testing strategies or validation they added alongside the templating solution.
Describe your experience with infrastructure-as-code tools (Terraform, Kubernetes, CloudFormation, etc.). How do you decide when to use Jsonnet versus native templating? Strong answers show nuanced judgment: Jsonnet for complex generation, native tools for simple cases. They mention specific scenarios where they chose Jsonnet and why.
How do you approach testing generated configurations? Walk me through a project where you validated that Jsonnet-generated manifests were correct. Excellent candidates describe testing strategies like schema validation, dry-run testing, or policy enforcement (OPA/Kyverno). They show that they treat configuration generation as seriously as application code testing.
Tell me about a Jsonnet library or framework you've built or contributed to. What design decisions did you make to ensure reusability? Look for evidence of thoughtful API design, clear abstractions, and consideration for downstream users. Strong candidates explain how they evolved their library based on usage patterns.
What's the most complex Jsonnet code you've written? What made it complex, and how did you keep it maintainable? Strong answers involve recursive functions, conditional logic, or large-scale parameterization. Candidates should discuss code organization, comments, and how they avoided clever-but-unmaintainable solutions.
Explain the difference between $ and $$ in Jsonnet. When would you use each? $ refers to the current object, $$ refers to the outer object. Strong answers explain variable scoping and demonstrate understanding of when you need to reference different object levels, particularly in nested structures or comprehensions.
How would you write a Jsonnet function that generates N copies of a Kubernetes Deployment with different replica counts? What parameters would you expose? This tests function design and parameterization. Strong answers create clean, understandable APIs for the function and show consideration for what parameters matter and which should be defaults.
Write a Jsonnet snippet that conditionally adds environment variables to a Kubernetes Pod spec only if a certain condition is true. Explain the syntax you used. Tests comprehension of conditionals and object merging. Strong answers use clean syntax and explain why they chose their approach (e.g., + operator for merging vs. full object reconstruction).
You have a Jsonnet file that's generating 10,000 Kubernetes manifests, but compilation is slow. How would you diagnose and fix the performance issue? Tests practical troubleshooting. Strong answers mention profiling, discuss algorithmic complexity (avoid O(n²) operations), and consider caching strategies. They show experience with large-scale generation.
Describe how you would organize a Jsonnet library to support multiple cloud providers (AWS, GCP, Azure) generating infrastructure manifests. What abstractions would you create? Tests architectural thinking. Strong answers show clear separation of concerns (cloud-agnostic vs. cloud-specific logic), reusable patterns, and how downstream teams would consume the library.
Build a Jsonnet template that generates Kubernetes Deployments for a microservices application. The template should accept parameters for: service name, docker image, replica count, and resource limits (CPU/memory). Generate at least 2 different services with different configurations from a single template.
Scoring rubric: (1) Template compiles cleanly (2) Parameterization is clear and well-named (3) Object structure is valid Kubernetes YAML (4) Demonstrates understanding of variable scoping (5) Code is readable and maintainable (6) Bonus: includes validation or error handling for invalid inputs.
Jsonnet developers are relatively specialized, so rates reflect mid-to-senior talent. These are 2026 LatAm market rates:
Typical US rates for comparison: Junior $60,000-80,000, Mid-level $100,000-140,000, Senior $150,000-200,000+, Staff $200,000-280,000+. LatAm talent offers 40-60% cost savings while maintaining quality, particularly for infrastructure specialization where experience matters more than fancy frameworks.
Rates are higher in Argentina and Mexico due to stronger local markets and demand. Brazil has the largest Jsonnet talent pool relative to cost. Specialized infrastructure expertise commands premium rates across all regions.
Latin America has built a strong infrastructure and DevOps engineering culture, particularly in Brazil and Argentina where companies like Globo, Mercado Libre, and enterprise tech firms have massive Kubernetes and infrastructure-as-code needs. These engineers work in UTC-3 to UTC-5 time zones, giving you 6-8 hours of real-time overlap with US East Coast teams. That overlap is critical for infrastructure work where you need immediate collaboration on deployments and incident response.
Jsonnet expertise in LatAm is concentrated but deep. Universities like USP (Brazil) and UBA (Argentina) produce engineers who naturally gravitate toward infrastructure specialization. Many LatAm developers have worked at regional scale-ups and multinational companies, giving them battle-tested experience managing thousands of resources across multiple environments. They're accustomed to complexity and systems thinking.
English proficiency among infrastructure engineers in LatAm is high, particularly among those with 3+ years of experience. Documentation, code comments, and architecture decisions benefit from clear communication. LatAm engineers tend to be thorough documenters, a critical skill when managing shared infrastructure libraries.
Cost efficiency is significant: a senior LatAm Jsonnet architect can cost $100,000-140,000 annually versus $220,000+ in the US. This isn't about cutting costs; it's about getting experienced infrastructure talent at sustainable rates. You get the same problem-solving rigor, just at better unit economics.
South's matching process for infrastructure specialists like Jsonnet developers focuses on three things: proven experience with configuration-as-code systems, a portfolio of work (open-source contributions or real projects), and demonstrated ability to communicate complex infrastructure decisions.
Here's how it works. You share your technical requirements: scale of configuration generation, which infrastructure tools you're using (Kubernetes, Terraform, etc.), and what's currently broken in your setup. South scours its pre-vetted network for developers with direct experience in those scenarios, not just generic DevOps experience.
Within days, you're interviewing developers who've actually built the systems you need. During the trial period, you'll evaluate not just coding ability but how well they document decisions and collaborate asynchronously. Infrastructure work requires strong communication, and South's vetting process tests that.
South offers a 30-day replacement guarantee. If the developer isn't delivering the quality or communication you need, we find a replacement at no additional cost. This applies to underperformance on the technical criteria you defined upfront, not personality clashes.
Ready to build a scalable configuration system? Start your search with South today.
Jsonnet generates JSON configuration files from templates. It's primarily used for Kubernetes manifests, Terraform variables, Prometheus alerting rules, and Grafana dashboards. Any scenario where you'd otherwise write hundreds of nearly-identical JSON or YAML files is a candidate for Jsonnet.
Jsonnet makes sense if you're managing multiple environments, have significant configuration duplication, or need dynamic resource generation based on input parameters. If your infrastructure is simple and static (few servers, few environments), plain YAML or Terraform might be sufficient.
Kustomize is lightweight and declarative, good for simple overlays. Helm packages applications with templating. Jsonnet is for deep configuration generation logic. Choose based on your use case: simple overlays (Kustomize), packaged apps (Helm), or complex parameterized generation (Jsonnet). Many teams use all three in different contexts.
Mid-level Jsonnet developers in LatAm range from $45,000-80,000/year depending on country and experience. Senior specialists command $80,000-140,000/year. This is 40-60% less than comparable US talent.
South can have qualified candidates in your pipeline within 2-3 days. The full hiring process from your initial conversation to signed offer typically takes 1-2 weeks, depending on your evaluation timeline.
It depends on your infrastructure complexity. For straightforward templating, a mid-level engineer can drive it. For designing reusable libraries or optimizing a large-scale system, you want senior experience. South can help you right-size the seniority level for your specific needs.
Yes. South places developers on flexible arrangements, including part-time contracts and project-based work. Be clear about project scope and timeline so South can find candidates available for exactly that duration.
Most South Jsonnet specialists are UTC-3 (Brazil, Argentina) or UTC-5 (Colombia, Peru). This gives 6-8 hours of overlap with US East Coast time (UTC-4/5) and 3-5 hours with US West Coast. Async work with morning standup meetings is the typical pattern.
South tests Jsonnet candidates with both practical coding exercises (building real templates) and architecture discussions (how they'd design large systems). We evaluate portfolio work, contributions to infrastructure projects, and their ability to explain complex decisions clearly.
South offers a 30-day replacement guarantee. If the developer isn't meeting your technical requirements or communication standards, we find a replacement at no extra cost. This applies to performance issues, not personality differences.
South handles all payroll, taxes, compliance, and equipment for team members in its network. You don't need to establish a legal entity in another country; South manages the administrative layer.
Absolutely. South can build out a full infrastructure platform engineering team for you, including Jsonnet specialists, Kubernetes engineers, and infrastructure architects. We've staffed entire DevOps teams for clients.
Strong infrastructure candidates have experience across the stack. Most South Jsonnet developers are proficient with Kubernetes, Terraform, Docker, and CI/CD platforms (GitHub Actions, GitLab CI). Many have AWS or GCP certification.
