Results from past projects, not a guarantee of future outcomes.
Start with the problem. You do not need to arrive knowing which technology to hire.
Cloud, SaaS and managed services with rising costs and little visibility into where the waste is.
Learn moreRepetitive processes, triage and hand-built integrations that AI and automation can take over.
Learn morePrototypes that never reach production and model spend with no control.
Learn moreBackend with no clear pattern, performance bottlenecks and difficulty scaling or evolving.
Learn moreManual or risky deployments, Kubernetes that is hard to operate and stalled migrations.
Learn moreWithout useful metrics, logs and alerts, problems show up for users first.
Learn moreA slow, poorly indexed site with no tracking to show what converts.
Learn moreAlternatives to managed services and licenses, evaluated by TCO before any migration.
Learn moreReal results from cloud infrastructure, migration and cost optimization projects.
A cross-cutting cost reduction and platform modernization program: PostgreSQL moved from Amazon RDS to StackGres, workloads moved from ECS to Kubernetes, Spot capacity with Karpenter and KEDA for demand-based scaling, lower network costs and a review of managed services. Consolidated program result: roughly US$18,000 per month in reduced infrastructure costs, while maintaining reliability and performance.
Consolidated TCP and HTTP service exposure behind an APISIX API gateway, removing the need for one NLB or ALB per service. What looked like a minor networking detail became a FinOps initiative: fewer load balancers, lower recurring network costs and a single point for routing, authentication and observability. Result: a 35% reduction in Load Balancer costs.
Evolved isolated monitoring into integrated observability, with an in-house stack of Prometheus, Grafana, Loki, Tempo, Mimir and Pyroscope centralizing metrics, logs, traces and profiling. Result: faster incident investigation, with signals correlated in one place and a 45% reduction in MTTR.
Migrated manual deployments and ad-hoc scripts to standardized GitOps pipelines in Azure DevOps, with reusable templates shared across projects, automated validation and predictable rollbacks. Result: a 40% reduction in deploy time and less pipeline configuration rework.
The engineering team spent excessive time analyzing logs and alert metrics to identify the root cause of incidents. I built an AI agent integrated with Slack and Datadog that analyzes logs and metrics in real time, suggesting fixes for recurring errors. Initial triage now happens in seconds, drastically reducing MTTR.
The main API had high latency during peak hours due to concurrency bottlenecks and inefficient database queries. I restructured the persistence layer, implemented distributed caching strategies and refactored critical flows for asynchronous processing. Average response time dropped from 1.2s to 180ms, supporting four times the concurrent requests on the same infrastructure.
Fixed-scope offers for anyone who wants a concrete first step.
Automate an operational process with AI.
Find where your technology operation can save money.
A website or landing page with SEO, Analytics and conversions set up.
Identify infrastructure risks, costs and bottlenecks.
A fixed-scope assessment on two fronts: operational efficiency (processes AI and automation can take over) and technology efficiency (cloud, Kubernetes, databases, observability, SaaS, managed services, open source alternatives and AI spend). Each opportunity becomes a business case with gain, effort, risk and priority, and whatever is prioritized can be implemented.
If your company is dealing with unstable deployments, backend that's hard to maintain, AI prototypes that never ship, or rising cloud costs without clarity, we can talk about an efficiency assessment.
A first conversation to understand the context and assess whether it makes sense to work together.
Six connected fronts, from AI and software to the cloud platform, reliability, cost reduction and digital growth. One-off projects or ongoing support.
Less manual work, AI in production
We automate processes that eat up your team's hours with AI and workflows, and put agents, RAG and LLM integrations into production, with observability and cost control per model, agent and workflow.
Learn moreFrom website to system, from MVP to scale
We build APIs, backend, frontend, internal systems, MVPs and digital products, with an architecture designed to scale and to be maintained. The focus is solving the business problem, not selling technical hours.
Learn moreInfrastructure that scales and ships fast
We modernize cloud, Kubernetes and delivery pipelines to cut deploy time, give teams autonomy and get the platform ready to grow, with infrastructure as code, GitOps and predictable rollback.
Learn moreVisibility to act before users notice
We create visibility into your operation to find failures faster and reduce downtime, with metrics, logs, traces, alerts and SLOs that help your team handle incidents with less effort.
Learn moreVisible costs, TCO under control
We reduce the total cost of running technology without sacrificing security, availability or the ability to innovate: cloud, Kubernetes, databases, SaaS, managed services and AI. Open source comes in when the TCO makes sense, not by default.
Learn moreA digital presence that converts
Building and modernizing websites and landing pages with technical SEO, performance and tracking set up from day one, so you can see which channels and actions actually generate leads and sales.
Learn morePrinciples that guide every project, from assessment to delivery.
Risks, decisions and trade-offs documented and shared with your team throughout the project.
Before-and-after metrics for every change: cost, deploy time, latency, MTTR.
Solutions that actually run in production, not just in a prototype or a slide deck.
The work starts with an efficiency assessment, with quick wins in the first weeks.
Documentation and mentoring so your team stays autonomous after delivery.
One-off projects or ongoing support, matched to each company's stage, without unnecessary complexity.
A practical approach to understand the current environment, prioritize risks and deliver sustainable improvements in infrastructure, backend and AI.
Understand the context, current pain points, technical goals, constraints and the maturity level of the team.
Assess infrastructure, cloud, Kubernetes, pipelines, observability, security, costs and operational processes.
Prioritize risks, improvements, quick wins and deliverables based on impact, urgency and complexity.
Implement improvements or support the team through execution with safety, documentation and governance.
Experience built across technology, retail, SaaS and international environments.
Calculators, checklists and practical guides on AI, Kubernetes, FinOps and platform engineering.
Terraform, Kubernetes EKS and observability on the infrastructure side; backend across multiple languages; LangChain, RAG and LLMs on the AI systems side.
I am Roberta Soliman, a Tech Lead working across Platform Engineering, software architecture and AI systems — from design to backend, from infrastructure to production operation.
I have experience architecting and operating critical platforms in AWS, GCP and Azure with Kubernetes, Terraform, CI/CD and observability; developing backend systems and APIs across multiple languages; and building AI systems with RAG and agents integrated with LLMs.
My work connects technical strategy, hands-on execution and production reliability, helping companies design solid architectures, reduce operational risk, and ship AI systems that actually reach production — not just a prototype.
My focus is to deliver practical, sustainable solutions aligned with each company's current stage, without unnecessary complexity.
Technical notes based on real infrastructure, backend and AI projects.
A practical guide to cutting cost and improving node scaling on Kubernetes with Karpenter, based on real projects.
NAT Gateway, cross-AZ traffic and EBS volumes: where the cost hides and what to fix first.
Estimate your monthly savings with Karpenter, Spot, off-hours shutdown and Graviton.
See what Spot really saves, after On-Demand fallback and Savings Plans.
Open source CLI that checks whether your Kubernetes workloads tolerate Spot interruptions.
See the toolOpen source CLI that checks manifests and images to find which Kubernetes workloads can run on arm64.
See the toolOpen source CLI that compares CPU and memory requests with observed usage (files or Prometheus) and tells you what to set.
See the toolPreflight, bake period, node and add-on order and safe rollback, in one page.
Right-size, scale on demand and use Spot safely, with results from real cases.
What the Agent Substrate, gVisor and OpenRL donation to CNCF means for agent platform governance, portability, security and production adoption.
Explore how Agent Substrate, gVisor, and OpenRL fit together through execution contracts, trusted evaluation, security, and Kubernetes operations.
Explore OpenKruise Agents, SandboxSet, SandboxClaim and warm pools, with Kubernetes examples and production guidance on security, observability and reliability.
Participation in events and initiatives such as DevOpsBootcampLive, AWS Members Day and BrazilClouds Talks, sharing hands-on experience in infrastructure, Kubernetes and service mesh.
Support for public-school students from Capão Redondo through the admissions process to international universities.
We evaluate managed services and proprietary tools, calculate the total cost of ownership and identify open source alternatives that genuinely save money, without trading license cost for operational complexity.
See Technology Cost OptimizationAlongside consulting, we maintain technical projects and open source initiatives focused on automation, infrastructure, backend and knowledge sharing.
A CLI that tells which Kubernetes workloads can safely run on Spot and what to fix in the rest. Reads manifests (Helm, kustomize) or the cluster, with table, JSON and CSV output, and a container image.
Prometheus exporter for Azure DevOps (Repos, Boards, Pipelines and Releases) with DORA metrics support. Kubernetes-ready, with a Helm chart, Grafana dashboard and alert rules included.
A support library for Jenkins pipelines in Windows/PowerShell environments, making automation, standardization and CI/CD step reuse easier.
An installation and setup guide for n8n with Docker, PostgreSQL and Slack integration, focused on workflow automation and integrations.
A fixed-scope assessment on two fronts: operational efficiency (processes AI and automation can take over) and technology efficiency (cloud, Kubernetes, databases, observability, SaaS, managed services, open source alternatives and AI spend). Each opportunity becomes a business case with gain, effort, risk and priority, and whatever is prioritized can be implemented. Deliverables: ai and automation opportunity map, cost analysis of cloud, infrastructure and saas, open source alternatives evaluated by tco, impact, effort, risk and payback estimates, 30/60/90-day roadmap and implementation plan.
01. Initial conversation · 02. Technical assessment · 03. Action plan · 04. Execution or ongoing support. A first conversation to understand the context and assess whether it makes sense to work together.
AI and Automation (Less manual work, AI in production); Software and Web (From website to system, from MVP to scale); Cloud and Platform Engineering (Infrastructure that scales and ships fast); Observability and SRE (Visibility to act before users notice); Technology Cost Optimization (Visible costs, TCO under control); Web Growth and Analytics (A digital presence that converts).
Yes. Building and modernizing websites and landing pages with technical SEO, performance and tracking set up from day one, so you can see which channels and actions actually generate leads and sales.
No. We reduce the total cost of running technology without sacrificing security, availability or the ability to innovate: cloud, Kubernetes, databases, SaaS, managed services and AI. Open source comes in when the TCO makes sense, not by default.
Work can start with a technical assessment and continue as one-off projects or ongoing support in DevOps, SRE, Platform Engineering, software architecture and AI engineering.
Clouds: AWS, GCP, Azure, Magalu Cloud. Terraform, Kubernetes EKS and observability on the infrastructure side; backend across multiple languages; LangChain, RAG and LLMs on the AI systems side.
São Paulo, Brazil · on-site in São Paulo and remote for Brazil and Latin America. Portuguese: native. Technical English for reading, documentation and asynchronous communication.
Book an initial conversation through Calendly, or reach out by WhatsApp, email or the contact form on this page. The first conversation is to understand your context and see whether working together makes sense.
The consultancy supports companies that want to cut manual work with AI and automation, reduce cloud, Kubernetes and SaaS costs, evaluate open source alternatives by TCO, evolve infrastructure and backend, and improve their digital presence with websites, SEO and Analytics. Work can start with an efficiency assessment and continue through focused projects or ongoing support in AI, DevOps, SRE, Platform Engineering, FinOps and software architecture.
