Active development
Full-stack mobile platform
Forge - Fitness Intelligence Platform
Active development
A production-style workout tracking platform with reliable set logging, progressive-overload recommendations, analytics, and a tested FastAPI backend.
Role and scope
Full-stack engineer across the Expo mobile app, typed client state, FastAPI services, PostgreSQL schema, migrations, test coverage, and Dockerized local development.
React NativeExpoTypeScriptFastAPIPostgreSQLSQLAlchemyAlembicDockerPytestTanStack QueryZustand
FastAPI + PostgreSQL
Backend
JWT + Argon2
Auth
Idempotent logging
Reliability
Case study
What this shows
Forge is a full-stack fitness intelligence platform focused on reliable workout logging, analytics, and recommendation workflows.
Problem
Workout tracking needs to stay reliable during active sessions, preserve set-level detail, and turn training history into useful recommendations without fragile mobile/backend state.
Product overview
Forge supports authenticated workout sessions, reusable templates, live set tracking, exercise history, bodyweight trends, estimated one-rep-max analytics, and progressive-overload recommendations.
Core user flow
Users register, create or reuse templates, run active workout sessions, log sets, complete or abandon sessions, and review history and recommendations.
Mobile architecture
The mobile layer uses Expo, React Native, TypeScript, TanStack Query, Zustand, SecureStore, Zod, and React Hook Form for typed data flow and ergonomic session state.
Backend architecture
FastAPI routes delegate to service classes, repositories isolate persistence logic, SQLAlchemy defines the relational schema, and Alembic manages migrations against PostgreSQL.
Reliability and offline behavior
Active workout drafts persist set details, input units, RPE, notes, set type, and rest timer timestamps. Queued mobile set logging and retry handling protect in-progress workouts from brittle network behavior.
Idempotency design
client_mutation_id values support idempotent set creation and backend reconciliation. Workout completion is also treated as an idempotent backend operation.
Analytics and recommendation engine
Forge includes working-set filtering, estimated one-rep-max calculations, bodyweight trends, previous-performance display, and rule-based progressive-overload recommendations.
Testing and CI
The repository includes backend and mobile tests, continuous integration, Docker-based PostgreSQL setup, seed data, documentation, and a Makefile.
Next milestones
Next work focuses on deeper analytics, richer recommendation quality, and continued reliability around active workout sessions.
Interested in the implementation?