Dedicated platforms and software architecture engineered for your niche

Rodrigo Pedruzzi — Software Developer and Quant.

I develop dedicated platforms and software architecture tailored to the specifics of each business, working across trading systems, SaaS applications and solutions for regulated data. Every project is built with technical rigor, information security and protection of sensitive data as structural requirements of the architecture.

4 systems in production 1 in closed pilot 2025—2026

Projects

In production

Casa Terapêutica Online

A web platform for remote child psychotherapy, based on digital sandplay, with collaborative construction of symbolic scenes inside a virtual house and real-time synchronization between therapist and patient. Generates observational reports in PDF from the scenes and clinical notes, using an LLM with prompt engineering to support different theoretical approaches.

Includes subscription and single-session billing, payments, coupons, licence management, an admin panel and usage metrics, plus infrastructure for scientific research with anonymized storage of scenes under consent.

A demo session in Casa Terapêutica Online: a symbolic scene built across the virtual house's six rooms, and an admin panel with access metrics, with no real patient data.

Stack: Flask · React · Supabase · WebSocket · Stripe
Period: 2026 — present
Role: Sole architect and developer
Scope: Public

View the product

In production

Agenda Clínica

A production scheduling platform for mental-health practices, managing recurring sessions, patient records with an evolution timeline, customizable services and drag-and-drop rescheduling.

Integrates a web app and a native Android app via Capacitor, with push and local notifications, using Supabase for authentication and data persistence and an admin panel for clinic management. Includes export in Excel and .ics formats and interoperability with Google Calendar, Outlook and iCal.

Agenda Clínica's monthly calendar and financial panel, in a demo environment with no patient data.

Stack: Vite · React · Supabase · Capacitor
Period: 2025 — present
Role: Sole architect and developer
Scope: Private

In production

Travessias Infantis

An editorial and interactive web platform for child psychology, integrating content, e-commerce, multimedia narrative and digital experiences for parents, children and mental-health professionals. Combines an editorial catalogue, a course, a blog and an interactive narrative with a soundtrack, visual effects and collectible elements, plus a Kids Mode with real-time multiplayer games, serverless synchronization and access via shareable links.

The infrastructure runs on static hosting and serverless services for database, authentication, real-time sync, edge functions and transactional email, keeping operation at zero infrastructure cost.

Travessias Infantis Kids Mode: memory game, chess, painting studio and free drawing.

Stack: React · Supabase Realtime · Render
Period: 2026 — present
Role: Sole architect and developer
Scope: Public

View the product

In production

Clinical Instruments Panel

A web platform to digitize and centralize the remote administration of assessment instruments, generating unique links and tokens, real-time tracking and a hub-and-spoke architecture where independent instruments integrate with a central panel and clinic services.

Built in Python and Streamlit, with Google Sheets via gspread as the persistence layer, it includes authentication, access control and integrations with scheduling, digital signature and document issuance.

Clinical Instruments Panel, with clinic name, patients and assessment instruments covered for confidentiality.

Stack: Python · Streamlit · Google Sheets (gspread)
Period: 2025 — present
Role: Sole architect and developer
Scope: Private

In development

Archivum

Archivum is a multitenant SaaS application for contract lifecycle automation, built by Perlaborar for businesses that handle standardized documents. The system parses .docx templates directly via OOXML, detects and classifies variable fields, and generates filled documents in .docx and PDF, with asynchronous processing. The architecture decouples the document engine from the web interface and uses Supabase for authentication, multitenant isolation, persistence and real-time sync.

Includes staged approval, electronic signature via Clicksign, third-party form filling, indexed search, document access control, reporting and payment integration via Stripe. The product is in the final stage of development, about to enter pilot testing.

Archivum's contracts panel and reports, showing the developer's own test data — no real contract or client shown.

Stack: Next.js · Supabase · Python (python-docx) · Docker
Period: 2026 — present
Role: Sole architect and developer
Scope: Private

Closed pilot

Quantitative research platform

A quantitative research platform for evaluating machine-learning meta-models in E-mini S&P 500 futures (CME), with order-by-order data ingestion, limit-order-book reconstruction and 45 causal microstructure features, subjected to automated non-leakage tests. The methodology combined Combinatorial Purged Cross-Validation, embargo, multiple-testing correction, Deflated Sharpe Ratio and a sealed out-of-sample confirmation set, with 111 registered trials and 260 automated tests.

The analysis concluded that the models showed no return statistically distinguishable from zero at the studied horizon, yielding reusable quantitative infrastructure and an evidence-based conclusion.

Real run of the 260-test automated suite (left) and a heatmap of mean t-stat per cell in the 0.2/0.3 grid, showing all 288 cells below the corrected bar (right).

Stack: Python · Order Book Reconstruction · Market Microstructure · Backtesting
Period: 2025 — 2026
Role: Sole researcher and developer
Scope: Public

View repository

About

I'm an independent software developer and quantitative developer, working across complex systems, modular architectures, backends, data pipelines, distributed systems and end-to-end SaaS applications. I also develop quantitative solutions for futures markets, focused on statistical research, backtesting, edge validation and risk architecture.

I work on software for healthcare and other regulated domains, including applications with sensitive data, authentication, payments, real-time features and privacy/LGPD compliance requirements, in collaboration with specialized professionals from each field.

I use high-performance AI-assisted development as part of the engineering process, pairing language models with technical validation, automated tests and architectural review to accelerate implementation, refactoring, testing and documentation without compromising quality, security or control over the system.

I work under contracts defined from requirements, deliverables and acceptance criteria established upfront. My core stack includes Python, C#, distributed systems, WebSocket, APIs and statistical analysis applied to financial time series. Professional written English and intermediate spoken English (B1).