Renan Machado Pinho — AI Engineer
AI Engineer
LLM Applications, RAG & AI Agents
Now
this month- ShippingLinkedifyLive
- BuildingNigma AIPre-launch
- WorkingCaterpillar Inc.
- BaseSão Carlos01:43

01 — About
About Me
I'm an AI Engineer with 5+ years of software engineering in banking and payments (Itaú, Santander, PagBank). Since 2025 I've been shipping LLM-powered products: RAG pipelines with LangChain, Claude and OpenAI at EcoAgro, AI agents with Mastra at Starsling, and two AI SaaS products of my own — Linkedify and Nigma AI.
I bring the discipline of high-throughput fintech systems (40,000+ requests per minute at Santander, 100K+ automatic debits per minute at Itaú) to LLM applications: asynchronous generation jobs, grounded and validated outputs, and pricing modeled on inference costs.
AI / LLM
RAG, AI agents & tool calling, LangChain, Mastra, Vercel AI SDK, pgvector, Whisper, GPT-5, Claude
Languages & backends
Python, TypeScript, Java, Go — FastAPI, Next.js, Elysia (Bun), Spring Boot
Cloud & infra
AWS (Lambda, ECS, S3, SQS, API Gateway), Cloudflare Workers, Docker, Kubernetes, Terraform, Coolify
Data & messaging
PostgreSQL (pgvector), MongoDB, AWS DynamoDB, Redis, BullMQ, Kafka, RabbitMQ, Drizzle ORM
02 — Products
Featured AI Products
Shipped end to end — model pipeline, product, billing, and infra.
Products
02
Live
01
Pre-launch
01
Last update
Oct 2026
Linkedify
SaaS that generates, schedules, and publishes technical LinkedIn posts with AI, conditioned on each user's topics and tone of voice.
GPT-5 + DALL-E 3 generation pipeline, a Vercel AI SDK tool-calling agent grounded with OpenAI web search, and async jobs on BullMQ/Redis behind an Elysia + Next.js + PostgreSQL stack. LinkedIn OAuth and Stripe tiers modeled on inference cost.
Linkedify Team
Marketing
The secret to our LinkedIn growth? Consistency. We post every weekday, same time — and engagement tripled in 3 months.

Summary
Supervised learning
Models learn from labeled examples to estimate outcomes for new data.
Nigma AI
For students and self-learners: generate questions and summaries from their own lectures, no matter where the class came from.
YouTube, recorded classes, PDFs, slides, and audio all go through the same pipeline — Whisper transcription, chunking, and OpenAI-generated summaries and review questions. Embeddings in PostgreSQL/pgvector for a planned chat-with-your-lecture feature. Backend on a Coolify VPS, frontend as a Cloudflare Worker. ~50 students and self-learners on the pre-launch waitlist.
03 — Skills
Technical Skills
AI & LLM
Languages
Cloud & infra
Architecture
Also working with
Specializing in LLM applications and the production systems behind them
04 — Pricing
Pricing
Daily
USD 250
for any task.
Full 8-hour working session.
Progress updates every 15 minutes with live demos.
Ideal for quick fixes, small features, and specific requests.
Payment: 60% before start, 40% on delivery.
Unused time is refunded — up to 35% back if tasks wrap early.
Crypto payments welcome (any token) — 5% off.
Monthly
USD 5000
to make your project live.
Every task delivered within a 48-hour window.
Sick days? You get that time back, no questions asked.
Built for shipping products from zero to production.
Payment: 40% upfront, 60% at the end of the month.
Comfortable working alongside existing teams.
Idle days are refunded — up to 35% back for unused time.
Crypto payments welcome (any token) — 5% off.
I am also available for hiring.
05 — Timeline
Career Journey
5+ years shipping production systems — LLM products since 2025, fintech at Itaú, Santander, and PagBank
Software Engineer
Contributing to a large-scale pricing system bridging legacy COBOL/mainframe environments with modern Java microservices.
Engineered robust file processing pipelines for high-volume structured data migrations, ensuring data integrity across the pricing lifecycle.
Built React.js frontend interfaces enabling internal teams to manage and monitor dealer file operations with greater visibility.
Leveraged AWS (S3, Lambda) for scalable, cloud-native file ingestion and processing workflows.
Software Engineer
Part of the Anti-Fraud team at one of Brazil's leading digital banks, building high-availability systems to detect and prevent fraudulent transactions in real time.
Delivered MVP 1 of the account and block management system in Java and Spring Boot, processing ~50K account blocks and unblocks per month.
Developed backend services for fraud detection rules and real-time transaction monitoring pipelines, handling ~40K requests per day.
Senior Full Stack Engineer
Architected RAG pipelines with LangChain and Python, integrating Anthropic Claude and OpenAI models for domain-specific retrieval and grounded generation.
Engineered the LLM application layer: document chunking, embedding generation, vector storage, retrieval, and prompt orchestration with structured outputs.
Built event-driven backend services with Node.js and the Serverless Framework on AWS for asynchronous, long-running generation jobs.
Led architecture decisions as technical reference for 2 teams, defining development guidelines with the PO, Tech Lead, and managers.
Senior Full Stack Engineer
Extended agentic capabilities with Mastra AI Agents and OpenAI, integrating them into onboarding and core product flows built on Next.js and TypeScript.
Updated the user onboarding flow to support member invitation.
Created seeding scripts and a local sandbox that mocks 3 third-party integrations (Supabase, Sentry, Slack), enabling offline development and reliable CI testing.
Software Engineer
Optimized 12+ Java and Spring Boot microservices, enabling the processing of 100K+ automatic debits per minute.
Increased ECS application performance by 55% by refining auto-scaling policies.
Designed a DynamoDB table purging process, reducing AWS costs by 10%.
Migrated AWS Lambda services from Python to Go, using Go's parallelism to double throughput from 200 to 400 SQS messages per Lambda instance.
Software Engineer
Built a messaging system using Java, Kafka, and RabbitMQ capable of processing 40,000+ requests per minute.
Implemented a retry mechanism for transactional email and SMS notifications.
Designed an event-driven system to track real-time operations and improve observability.
Delivered hundreds of automated tests using JUnit and Mockito ensuring code quality.
Participated in MVP 1 delivery for credit lines of global cards.
Authored technical documentation and introduced branch/release models for the team.
Full Stack Software Engineer
Developed backend applications using Java 11, Spring Boot, JSF, and Oracle Database.
Fixed key bugs, increasing sales productivity by 10%.
Optimized memory consumption for large-scale report generation, reducing overhead by 50%.
Created a logistics QR Code and barcode printing service for internal operations.
Full Stack Software Engineer
Developed a hearing aid sales platform using React and Node.js.
Deployed the company's first MVPs for internal and external platforms.
Integrated Google Analytics and ad campaigns, growing the platform to 10,000+ users.
Led deployment of MongoDB solutions for scalable data management.
Senior Full Stack Software Engineer
Developed a fitness center application using Next.js and Supabase for workout management, scheduling, and student performance tracking.
Contributed to an e-learning platform with AWS and Java, facilitating content distribution to thousands of users.
Built full-stack CRM applications using TypeScript (Node.js) for small local businesses over 2+ years.
Implemented deployment pipelines for frontend applications with optimized CI/CD configuration.
Trainee
Worked on the maintenance of e-commerce applications created in WordPress.
Maintained WordPress plugins and supported bug fixes.
Contributed to day-to-day platform stability and team delivery.
Key Technologies & Expertise
06 — Writing
Technical Insights
I write about LLM applications, RAG pipelines, and the production systems behind them — plus lessons from shipping AI products and high-throughput fintech services.
LLM Applications
RAG, agents & structured outputs
Production Systems
Async jobs, evals & infra
Architecture
Reliable systems at scale
07 — Contact
Get In Touch
Let's build something amazing together
© 2026 Renan Machado Pinho. Built with React, TypeScript & Framer Motion.