Alisson Rosa
Principal Engineer, Data & AI | Agentic Architect | Google Developer Expert (Cloud & AI)
Belo Horizonte, MG, Brazil (Global / Remote)
Professional Summary
Principal Engineer, Data & AI Architect, and Technical Leader with extensive experience designing cloud-native data platforms, distributed backends, and production agentic AI systems. Recognized by Google as a Google Developer Expert (GDE) in Cloud & AI and active contributor in Gemini early-access product feedback programs. Track record includes cutting Fintech Data Lake costs by 88% (Azure and Databricks to BigQuery), cutting Databricks spend by 68% with FinOps cost segmentation, working on Latin America's largest big data migration to GCP (40 PB) alongside Google Professional Services (PSO), and building from scratch the Americas' largest livestock Data Lake (AWS to GCP) that catalyzed an acquisition by a European group.
Core Competencies & Technical Skills
Programming & Distributed Systems: Go (Golang), Python, Rust, SQL, gRPC, Protocol Buffers (Protobuf), Optimistic Concurrency Control, Distributed Scheduling, Microservices, REST APIs, Linux.
Agentic AI & Machine Learning: Google Agent Development Kit (ADK Go / ADK Python), Gemini Models & API, Vertex AI, Model Context Protocol (MCP), Agent-to-Agent (A2A), Human-in-the-Loop (HITL) Workflows, LLM Orchestration, RAG.
Data Engineering & Analytics Platforms: BigQuery, Databricks, Unity Catalog, OpenMetadata, Apache Airflow, Dexaflow, Apache Spark, Dataproc, Hadoop, Oracle, Greenplum, Greenfield Data Lakes, Data Warehouse, Master Data Management (MDM), Dataform, dbt, Business Intelligence (BI), ClickHouse, PostgreSQL.
Cloud Infrastructure, FinOps & Leadership: Google Cloud Platform (GCP), Amazon Web Services (AWS), Cloud Run, Cloud Firestore, Eventarc, Cloud Tasks, Kubernetes (GKE), Docker, DevOps, CI/CD, Cloud FinOps, Engineering Management, Technical Leadership, Agile Delivery.
Selected Enterprise Case Studies & Measurable Impact
88% Data Lake Cost Reduction (Azure and Databricks to BigQuery)
Architected and led the migration of an enterprise Data Lake from Azure and Databricks to Google BigQuery at a Brazilian Fintech, achieving an 88% reduction in infrastructure cost while improving query SLA and data governance.
68% Databricks Cost Reduction with FinOps
Cut Databricks cost by 68% on a multi-tenant data platform by rolling out FinOps with per-tenant cost segmentation, while modernizing it with AWS DMS, S3, Databricks Serverless, dbt and Airflow.
40 PB Big Data Migration with Google PSO
As a partner engineer with Google Cloud Professional Services (PSO), worked on the largest big data migration in Latin America, moving 40 PB from Hadoop to Google Cloud with Dataproc, Dataflow, Data Fusion, BigQuery and Pub/Sub, then on further PSO engagements in Brazil.
Americas' Largest Livestock Data Lake, then a European Acquisition
Founded and architected from scratch an AgTech Data Lake on AWS and subsequently led its migration to GCP, establishing the largest livestock data platform in the Americas, driving revenue growth and enabling the company's acquisition by a European group.
Experience
FinOps assessment targeting 30% lower multicloud spend; client allocation that cut Databricks cost by 68% with cost segmentation.
Data platform maturity assessment, BigQuery editions migration study and real-time architecture design.
Built the data engineering team; cut data lake cost by 88% moving from Azure and Databricks to GCP and BigQuery; real-time data for operations.
Data lake and lakehouse architectures for clients in several industries; led MLOps teams and mentored data engineers.
Big data solutions on Hadoop and Google Cloud, including a large migration program delivered with Google Cloud Professional Services.
Designed the data architecture and led the AWS to Google Cloud migration of the company's livestock data platform.
IoT analytics architecture, data warehouse and data lake, on-premises Hadoop cluster.
Open-Source & Systems Architecture
Firestore Session Service for Google ADK Go v2
Proposed a Firestore-backed session service, google.golang.org/adk/session/firestore (PR #2058 under review, issue #2057), for durable HITL agent pause and resume on serverless GCP stacks, with optimistic concurrency control and 91.4% test coverage.
Data Catalog Governance, Unity Catalog & Airflow 3
open-metadata/OpenMetadata: reported (#32063) and fixed (PR #32065, merged) a login dropped when the ServiceWorker fails; reported the Unity Catalog BROWSE-only ingestion gap (#33505). apache/airflow: reported two Airflow 3 UI bugs fixed upstream within a week (#67883, #67884).
Dexaflow: Airflow-Compatible Orchestrator in Go
Re-engineered Apache Airflow's scheduling and REST/UI control plane into a single Go binary, eliminating Python GIL bottlenecks while providing Kubernetes pod-per-task isolation and GitOps-native DAGs.
Awards, Community Speaking & Languages
Google Developer Expert (Cloud & AI) & Gemini Feedback
Official Google Developer Expert in Cloud & AI and active contributor in Google's early-access Gemini product feedback loops.
TDC · DevFest BH · Data Science Academy
Speaker at The Developers Conference (TDC) and DevFest BH; Organizer & Speaker at Data Analytics Meetup; Technical Mentor at Data Science Academy.
Portuguese & English
Portuguese: Native
English: Intermediate (B2, Professional Technical Working Proficiency)