Projects & Experience — Danrley Pereira

AI Engineer — DW Corp

2022-09 to present

My company — LLM agents and reusable product foundations.

  • Built LangChain/LangGraph agents (a WhatsApp budget assistant and an event-driven multi-agent bus) behind a common multi-provider LLM interface.
  • Built Curupira, a reusable Rust/OIDC authentication platform (multi-tenancy, SSO, MFA) that cuts new-product setup time.
  • Set up reusable foundations — payment integrations, CI/CD, and a bare-metal + hybrid-cloud stack with Terraform and Grafana observability.
  • Hired and mentored engineers, and lead Brasília's first Software Craftsmanship community.

Technologies: Python, LangGraph, Rust, React, Terraform

Researcher — Ibict / MCTI

2025-03 to present

Data pipelines and RAG research at a federal institute.

  • Built Snowflake data pipelines over institutional data, driving progress toward training ML models on proprietary data.
  • Co-advised an undergraduate thesis that designed a RAG system for the Brazilian judiciary (graded 9.0/10).
  • Ran a technical audit of a 600K+-user platform against ISO/IEC 25000 and LGPD, producing a prioritized remediation roadmap.

Technologies: Python, Snowflake, RAG, FastAPI

AI Engineer — Epicor

2023-11 to 2025-02

AI-assisted procurement automation connected to the Epicor ERP.

  • Integrated Azure OpenAI and a RAG pipeline over proprietary data; used Azure Document Intelligence for structured extraction from supplier documents.
  • Designed .NET Aspire microservices with RabbitMQ event-driven processing, deployed on Kubernetes.
  • Led an Angular integration portal refactor into microfrontends, with strong xUnit and end-to-end test coverage.

Technologies: .NET, Azure OpenAI, RabbitMQ, Kubernetes, Angular

Senior Full-Stack Consultant — SubSurface Dynamics

2024-01 to 2024-06

Real-time operational monitoring for thousands of users.

  • Refactored a MERN app into a modern TypeScript project on Fastify, Socket.io, Redis and Typegoose.
  • Scaled real-time streaming with Redis and tuned MongoDB Atlas time-series for cost and performance.
  • Secured the platform with Azure AD (ADAL) + Auth0 (OAuth2/SAML); added dark mode and a notification system.
  • Built 2D/3D charts and a Three.js wellbore visualization to support real-time decisions.

Technologies: TypeScript, Socket.io, Redis, React, Three.js

Personal Project — jackBlack

2024-01 to 2025-06

Deep RL agent that learns Blackjack by self-play.

  • Trained a Deep Q-Network (TensorFlow/Keras) with experience replay and a target network.
  • Engineered an 8-deck shoe simulator with Hi-Opt II card-counting encoded into the state.
  • Ran GPU training on NVIDIA (CUDA) and AMD (ROCm) with iterative fine-tuning.

Technologies: TensorFlow, Python, CUDA, ROCm

Data & DevOps Engineer — Dado Global

2021-12 to 2023-04

Python ETL platform; improved delivery SLA from 70% to 95%.

  • Designed an ETL system with layered responsibilities; SQLAlchemy with rigorous pytest coverage of the transform/load layers.
  • Automated the extraction layer with Selenium and Behave/Gherkin.
  • Moved delivery to Continuous Delivery on GitHub Actions (weekly → daily), lifting the SLA from 70% to 95%.

Technologies: Python, SQLAlchemy, Pytest, GitHub Actions

Full-Stack Engineer — Compuletra

2020-06 to 2021-08

Logistics: transport, delivery and yard management.

  • Built transport, delivery and yard-management software focused on truck and fleet operations.
  • Led a vehicle-inspection management system and an ERP desktop platform (Electron) on PostgreSQL.
  • Built a commercial page generator with a PCI-compliant checkout; plans updated from the database without redeploys.

Technologies: C#/.NET, React, Electron, PostgreSQL

Platform Developer — Ministry of Health / USP

2019-06 to 2020-07

Medicine distribution across 5,000+ municipalities.

  • Helped the Ministry of Health's pharmaceutical department manage medicine distribution across 5,000+ municipalities.
  • Built the frontend and integrated REST APIs and microservices to surface data and simplify task execution.
  • Reduced fraud and streamlined service integration at national scale.

Technologies: Angular, Python, Flask, PostgreSQL

Software Engineer — ClearSale

2021-11 to 2022-03

High-performance anti-fraud microservices.

  • Built and tested anti-fraud microservices in C#/.NET; optimized large-scale queries with Dapper.
  • Designed DB views, triggers and stored procedures for real-time detection and alerting.

Technologies: C#/.NET, ASP.NET, Dapper, SQL Server, Azure Service Bus

Technical Lead & Researcher — LabTech, UDF

2023-01 to 2023-12

Multi-tier event-management platform.

  • Led a multi-tier platform (React frontend, Python logic, Spring Boot data) with RabbitMQ and Kubernetes.
  • Organized teams by technology and mentored young professionals into the industry.

Technologies: React, Python, Spring Boot, RabbitMQ, Kubernetes

Software Engineer — Layers Education

2022-03 to 2022-09

Backend & microfrontends; performance and security.

  • Maintained backend services and white-label microfrontends; cut MongoDB query times from 1.3s to 0.6s.
  • Built a BigQuery data lake for ETL and hardened security to OWASP standards.

Technologies: Node.js, Vue, MongoDB, BigQuery, GCP

Full-Stack Developer & BPM — ConsultMídia / Bank of Brasília

2020-04 to 2021-01

Workflow modernization for a bank (Camunda BPM).

  • Built AngularJS/Spring Boot interfaces and REST APIs on Camunda to modernize the Bank of Brasília's workflows.
  • Integrated Alfresco (document management) and automated performance tests (JMeter, Selenium).

Technologies: AngularJS, Java, Spring Boot, Camunda, Alfresco

Full-Stack & Mobile Developer — Transoft

2021-02 to 2021-10

Fleet sensor/camera and staff-management apps.

  • Revamped an application managing sensors and cameras for bus fleets (Angular, Laravel).
  • Built a mobile staff-management app integrated with Oracle (Ionic, Angular).

Technologies: Angular, Ionic, Laravel, PHP, Oracle