Skip to main content
Accenture Nordics · Oslo · Norway

Fábio
Rodrigues
Pereira.

Team & Tech Lead · Senior Software Engineer · Quant Researcher

I lead a cross-functional team at Accenture Nordics, owning 100+ microservices in a high-volume Norwegian financial system. On the side, I build production RL trading agents on the Brazilian B3 exchange and ship SaaS suites running in real businesses.

GitHub ↗ LinkedIn ↗ Download CV .PDF Say hi
Photo of Fábio Rodrigues Pereira, Team and Tech Lead & Senior Software Engineer
PORTRAIT · 2026
REC ●
OSLO · NO 29.05.1988
fabio.rp Online
Scroll
02 / ABOUT

Team & Tech Lead in a mission-critical system.
Quant researcher by obsession.

I'm a Brazilian–Italian engineer based in Oslo, working at the seam between production product engineering and quantitative trading research. By day I lead a team operating one of the Nordics' largest financial transformations. By night and weekend I train PPO agents on six years of B3 tick data.

My weeks rotate between Java/Spring Boot microservices, Kafka pipelines, Python scripts, and MT5 algorithm live trading. I care about systems that stay on after the demo — blameless postmortems, walk-forward CV, paper-trading dashboards, Telegram heartbeats, and boring-but-correct double-entry accounting.

Based
Oslo, Norway
Experience
5+ years
Education
MSc · UiO
Languages
PT · EN · NO · ES
// stack
v2026.05
Languages 06
JavaPythonKotlinTypeScriptSQLC/C++
Backend 05
Spring BootHibernateKafkaRESTSOAP
Frontend 03
ReactAstroNext.js
ML / RL 04
PyTorchTensorFlowTransformersOptuna
Infra 07
DockerOpenShiftJenkinsAzureGCPSQLServerMT5
Ops / SRE 05
SplunkCamundaJiraITILGrafana
03 / EXPERIENCE

Current role &
career path.

Team & Tech Lead at Accenture Nordics — production-grade Java microservices for one of the largest Nordic financial transformations. Trading research runs in parallel since 2011.

Current · Featured Sep 2022 — Present

Team Lead | Tech Lead & Senior Software Engineer

Accenture Nordics · Oslo, Norway

Lead a cross-functional team in the Managed Services unit for the company's largest client — the largest Nordic IT transformation project in history (Konsultprisen 2025 winner in “Store oppdrag”).

  • Promoted to Team Lead | Tech Lead in Jun 2025, still hands-on as senior engineer.
  • Lead 1 functional analyst, 1 data expert, 1 tester and 6 developers — ensure high-quality delivery and continuous improvement.
  • Own production operations of 100+ microservices in a high-volume financial system (300k+ active users).
  • Drive ITIL-aligned incidents/problems/changes, blameless postmortems, on-call response.
  • Architect React/TypeScript frontend, Spring Boot + Hibernate microservices, and Kafka event-streaming pipelines.
  • Design and improve CI/CD pipelines (Jenkins, OpenShift, Azure, Bitbucket).
  • Lead an internal MCP (Model Context Protocol) server on Spring AI / Spring Boot, integrated into the team’s IntelliJ AI assistant chat — automating managed-service operations across the retirement-pension and AFP product lines for 200+ active internal users.
Scope · today
Software Engineer since
Sep 2022
Team & Tech Lead since
Jun 2025
Microservices
100+
Active users served
300k+
Team size
9
Key technologies
JavaSpring BootSpring AIMCPHibernateKafkaReactTypeScriptOpenShiftJenkinsAzureSplunkCamunda
Earlier experience
  1. 2011 — Present

    Financial AI Strategist

    Independent research · Remote

    Develop algorithmic trading, forecasting, and portfolio strategies for US, Brazilian, and European markets. Deep RL agent (T70) in MT5 production.

    PythonPyTorchTensorFlowOptunapandasNumPyMetaTrader 5
  2. Aug 2017 — May 2018

    Team Lead | Basketball Coach

    Oslo Studentenes Idrettsklubb (OSI) · Oslo, Norway

    Team Manager for the H3 basketball squad, third division of the Norwegian Senior Championship. Tactical coaching and strategic direction over the 2017/18 season.

  3. Aug 2012 — Nov 2014

    Lawyer & Accredited Public Defender

    Defensoria Pública do Estado de São Paulo · São Paulo, Brazil

    Civil and criminal public litigation portfolios as independent attorney and accredited public defender. Legal consulting for a notable electronics distributor in Brazil.

Education
  1. May 2022
    MSc Computational Science: Applied Math & Risk Analysis
    Universitetet i Oslo
  2. Jun 2019
    BSc Mathematics and Economics (Minor: Finance, Insurance & Risk)
    Universitetet i Oslo
  3. Dec 2014
    Post-graduation in Corporate Labor Law
    Complexo Damásio, São Paulo
  4. Jun 2012
    BSc Law
    Mackenzie University, São Paulo
★ FEATURED · 04 LIVE SINCE MAY 2026

T70 Reinforcement-Learning Trader

PPO agent trading WIN mini-index futures on B3 (Brazilian Mercantile Exchange). Walk-forward backtested across 19 folds and 6 years of tick data. In production via MT5 since May 2026.

Cumulative P&L · Backtest

Jul 2021 → Mar 2026

977 trading days
1,894 trades
Equity curve Max DD +R$ 340,919
Performance · Net
v3 ensemble
Sharpe
1.70
CAGR
46.62%
Sortino 2.49
Calmar 1.02
Total P&L (net) R$ 340,919
Max drawdown -R$ 45,555
Win rate 46.04%
Profit factor 1.18
Payoff ratio 1.39
Avg win +R$ 2,527
Avg loss -R$ 1,823
Trades 1,894
Trading days 977
Trades / day 1.94
Avg duration 2.7 bars
Time in market 2.54%

PPO policy with TP/SL plus limit-order entries, trained via walk-forward CV over 19 folds on 6 years of tick data. Production deployment: T70 ensemble in MT5 since May 2026, paper-traded daily with Telegram heartbeat. Currently being superseded by v4 — a 6-advisor reward-diverse ensemble.

Python PPO stable-baselines3 MetaTrader 5 Optuna Azure
05 / WORK

Selected
projects.

An AI-generated animated series, two SaaS suites running in real businesses, and the marketing operation that keeps a clinical practice growing.

5.1 — AI Pipeline

Equipe Coral

Children's animated series produced end-to-end with an AI pipeline — script, scenes, render. 70MB+ of HTML and video assets generated via prompt-driven analysis and ffmpeg compositing.

  • 01Gemini drafts script + scene metadata
  • 02HTML scenes generated and styled procedurally
  • 03ffmpeg renders to FCPXML for final cut
Gemini ffmpeg FCPXML HTML
Watch on YouTube ↗

▶ Latest episode · youtu.be/sSztMip9NJE

Hotel Diretor backoffice home — quick-access modules: Reservas, Mapa de Reservas, Folio, Housekeeping, Conversas, Quartos, Notas Fiscais, Contas a Pagar, Inventário
5.2 — SaaS Suite

Hotel Diretor

Hotel operations from check-in to in-room TV to WhatsApp guest support — one integrated suite. Three products, one source of truth.

  • PMSProperty management system core — Java backend
  • TVAndroid-based digital signage for hotel TVs — Kotlin
  • BOTWhatsApp CRM & guest bot — Java + Twilio + Postgres
Java Kotlin TypeScript Postgres WhatsApp API
Visit hoteldiretor.com.br ↗
Restaurante Diretor backoffice home — quick-access modules: POS, Ordens, Mesas, Salão, Reservas, Itens, Produtos, Caixa, Relatórios
5.3 — SaaS Suite

Restaurante Diretor

Restaurant POS suite — from kitchen display to backoffice to manager portal. Four products coordinating one service.

  • BOBackoffice — Java admin backend
  • UIPortal — manager web UI, TS/JS
  • KDSKDS App — Android, Kotlin
  • OPSOps — operations tooling
Java Kotlin TypeScript Postgres Android
Visit restaurantediretor.com.br ↗
5.4 — Web + Marketing

MinhaAnalista

End-to-end web presence for a clinical psychology practice — site, SEO, paid ads, and analytics, all managed in one hand. Live at minhaanalista.com.br.

  • SITEAstro + React + TypeScript
  • SEOTechnical SEO + sitemap + Search Console
  • ADSGoogle Ads paid search campaigns
  • DATAGoogle Analytics performance dashboards
Astro React TypeScript MUI Bootstrap 5 Google Ads GA4 Search Console
Visit site ↗
06 / RESEARCH

Publications.

MSc thesis

Trading Financial Markets Using Reinforcement Learning: Application and Analysis

· Universitetet i Oslo · MSc Computational Science

Operational-capital-cap PPO agents trading Brazilian equity futures, walk-forward backtested across 19 folds and 6 years of WIN mini-index tick data.

Reinforcement LearningQuantitative FinancePPOB3
View on GitHub ↗
  1. Best paper

    WNNLP 2021 · Best Scientific Academic Article
    NLPTransformersMultitask Learning
  2. Paper

    FYS-STK4155 · Applied Data Analysis & ML, UiO
    RNNLSTMTime Series
  3. Paper

    FYS-STK4155 · UiO
    Deep LearningRegressionClassification
  4. Paper

    FYS-STK4155 · UiO
    RegressionNumerical Methods
07 / CONTACT

Say hi.

Questions, project ideas, or just a chat about RL, Java microservices, or trading the B3 — the inbox is open.

fabiorodp@gmail.com GitHub ↗ LinkedIn ↗ Download CV .PDF