Backend & APIs
REST services that hold up in production: authentication, multi-tenant data, integrations, background jobs and observability.
- Python
- FastAPI
- Flask
- Chalice
- Laravel
Curitiba, Brazil · remote has no borders 🌎
Alisson Pelizaro.
Senior Software Engineer · TOTVS · Founder of Element
15+ years turning ideas into clean, scalable code. Today I build generative AI — LLMs, RAG and agents — on top of Python backends and cloud-native architecture. And I run my own software house.
- 0years of career
- 0companies — state-owned to big tech
- 0clouds (GCP · AWS · Azure)
15
Ap
Alisson Pelizaro
python · llm · cloud
Python◆FastAPI◆Flask◆LangChain◆
LlamaIndex◆RAG◆OpenAI◆PostgreSQL◆
Redis◆Docker◆Terraform◆GCP◆
AWS◆Azure◆Vue.js◆Angular◆
Next.js◆Node.js◆Laravel◆PyTest◆
GitHub Actions◆
What I do
Full stack, with the backend as home.
I like problems that involve APIs, automation and something clever with AI. Below is what I reach for every week — not a list of things I once touched.
Generative AI
LLMs inside the product, not as a demo: RAG over real documents, agent workflows, natural-language interfaces and evaluation that actually measures something.
- LLM
- RAG
- Agents
- OpenAI
- LangChain
- LlamaIndex
Cloud & DevOps
Infrastructure as code, containers and pipelines that ship on merge. Three clouds in production, GCP as the daily one.
- GCP
- AWS
- Azure
- Docker
- Terraform
- GitHub Actions
Frontend
Interfaces built with the API on the other side in mind — the contract between the two is never an afterthought.
- Vue.js
- Angular
- Next.js
- Node.js
- TypeScript
Data
Relational, key-value and document stores. Modelling, indexing and query plans — because the slow part is almost always the database.
- PostgreSQL
- MySQL
- Redis
- DynamoDB
Quality & mentoring
TDD, code review and CI that blocks the bad merge. Plus the part I enjoy most: mentoring, knowledge sharing and building community.
- PyTest
- PHPUnit
- Robot Framework
- TDD
Selected work
Products in production, with real users.
Built through Element, my software house — product, engineering and infrastructure under one roof.
The journey
From a first HTML page to production AI.
The first site
Built and published my first website. Nothing about it was good — but it was live, and that was enough to decide the next twenty years.
Apprentice running all of IT
An HR company in Curitiba handed a teenager its entire IT and digital marketing. In parallel, technical courses in Linux and web design.
Systems analyst, ATMs
Six years and nine months reading Java stack traces over SSH to find why an ATM had failed somewhere in southern Brazil. Log analysis, Linux, VBA — and the internal web tools that automated my own team.
Degree, and a promotion
Graduated in Systems Analysis and Development, and moved up to Advanced Support Analyst. Technical courses in operating systems and C++ along the way.
I started selling my own work
Opened a web design and digital marketing company, and switched roles at Caixa to specialised support and PHP development. The first in-house CRM, from scratch and with no framework, came out of this.
PHP, Asterisk and real clients
Two years building web and desktop applications over RealTime Asterisk servers — PBX management. On the side: Blanche Turismo, Pizzaria Treviso, Legis Dei and Portal RC.
SaaS, at scale
Two and a half years on a business management SaaS, mid-level to senior. PHP and Laravel, Python, Vue and Node on the front, and AWS underneath — SQS, S3, API Gateway, SES, DynamoDB.
Senior software engineer
Generative AI as the daily job: LLMs, RAG and agents on Python backends, Angular and Next.js on top, GCP underneath. In parallel, two years at outsera on AI solutions.
My own software house
I own and run Element: a Brazilian software house shipping mobile, desktop and web products end to end. Monnif, Postgly, GS Piscinas and Falcão Lage came out of it.
Where I work
Large-scale software, at the biggest software company in Brazil.
TOTVS · TOTVS Labs
Senior Software Engineer · since 2022
I am a Senior Software Engineer at TOTVS, working on generative AI inside products that thousands of companies depend on at the same time. At this size the engineering changes: multi-tenant is the default, cost per request is a real budget line, and a change that looks small in the editor can be expensive in production.
TOTVS is the largest software company in Brazil and the leading ERP vendor in Latin America. Founded in 1983 and publicly traded on B3 (TOTS3), it serves companies across practically every sector of the Brazilian economy — retail, agribusiness, manufacturing, healthcare, services. Software at that reach is infrastructure: when it stops, it is never one company that feels it.
- Founded
- 1983
- Listed on
- B3 · TOTS3
- Market
- ERP leader in Latin America
- My focus
- Generative AI · Python · GCP
- Multi-tenant by defaultA bug does not hit one user, it hits every client on the tenant. Data isolation and blast radius belong in the design, not in the post-mortem.
- Scale changes the mathsAt volume, milliseconds and fractions of a cent per request turn into budget. Performance stops being taste and becomes cost.
- AI in production, not in a demoAn LLM inside an ERP means evaluation, guardrails, cost per token and a fallback for when the model is wrong. Anyone can ship the demo.
Let's talk?
A project, a role, a question about AI in production, or just to say hi. I read everything and I answer.