Software Engineer · AI · Cloud · DevSecOps
i build ai systems —
the rest, automated.
End-to-end systems at the intersection of DevSecOps on AWS, RAG / LangChain and applied AI. From cloud infrastructure to intelligent applications that survive real users. São Paulo, Brazil · Available for remote projects.
about
Marcelo Manara
São Paulo, Brazil
A little about me
Software Engineer in constant evolution — building at the intersection of Cloud infrastructure, AI systems, and applied GenAI.
Currently working as an AI Engineer at a law firm, developing document intelligence engines with Python, PostgreSQL, RAG and LangChain — automating legal document analysis and reducing manual triage time significantly.
Previously a Cloud & DevSecOps Intern at Compass UOL, where I designed scalable AWS architectures, automated CI/CD pipelines with Jenkins, Docker and Kubernetes, and managed Linux infrastructure following DevSecOps and FinOps practices.
I believe the most impactful solutions live at the intersection of Cloud, security and intelligent data. Advanced English and always building.
experience
- Building AI automation workflows with Python, PostgreSQL and LangChain.
- Developing a document intelligence engine for parsing and extracting data from legal PDFs.
- Implementing AI-driven business rules to assess the viability of legal appeals — reducing manual triage time significantly.
- Designed scalable and secure cloud infrastructures on AWS via IaC (EC2, RDS, VPC, ALB, Auto Scaling).
- Automated full CI/CD pipelines with Jenkins, Docker and Kubernetes via GitHub webhooks.
- Managed Linux servers following DevSecOps and FinOps best practices.
skills
Hover over the belt to pause
lab
selected workSINAPSE — B2B LegalTech
SaaS platform for law firms with a document intelligence engine. Semantic cross-RAG search with real-time responses and zero hallucination.
AI Document Extraction Pipeline
End-to-end pipeline that ingests documents and extracts structured data with LLMs — parsing, chunking, embeddings and schema validation in one automated flow.
AI Appeals Analysis Platform
Job-oriented backend with a durable PostgreSQL-backed queue (no external broker), specialized domain workers, feature-flag migration control, and automatic reconciliation of stuck jobs.
Deterministic Legal Cost Calculator
A rules engine where AI authors rule candidates but never touches the calculation path. AST static validation gates promotion; production math is deterministic, versioned, and fully auditable with Decimal precision.
AI Critical Publication Analysis
Database-first pipeline classifying high-volume unstructured text, with multi-provider LLM fallback, token/latency observability, and automated executive PDF reporting.
AI Document Capture Pipeline
End-to-end document lifecycle orchestration — download, AI analysis, publication, retention cleanup, and self-healing reconciliation, with database-first deduplication and per-stage observability counters.
Scalable WordPress on AWS
High-availability infrastructure with EC2, RDS, EFS and Auto Scaling. Architecture focused on resilience and cost-efficiency.
CI/CD Pipeline with Kubernetes
Jenkins pipeline that automates the build and deploy of a FastAPI application to a local Kubernetes cluster via webhooks.
products
shipping soonSINAPSE
AI-powered LegalTech SaaS turning static legal PDFs into a searchable intelligence engine — RAG with hard grounding, automated metadata extraction and permanent semantic memory.
contact
Available for projects, collaborations and remote opportunities.