Projects

Projects

7 projects delivered to production — automation, AI, and enterprise integrations.

01

HYPR AI Performance Review

IAApps ScriptAutomation

AI-powered performance review system for 150+ employees. Automatic generation of structured feedback and objective scoring.

Problem

Manual performance reviews took 3 weeks per cycle and suffered from significant subjective biases, reducing employee trust.

Stack

Google Apps ScriptOpenAI APIGoogle SheetsGoogle Forms

Results

  • →150+ employees evaluated per cycle
  • →Review time reduced from 3 weeks to 3 days
  • →Manager satisfaction score went from 62% to 91%
GitHub →
02

BigPicture Initiative Tracker

JiraApps ScriptIntegration

Cross-platform initiative tracking synchronizing Jira, Confluence, and Google Sheets for real-time project progress visibility.

Problem

Teams used 3 different tools with no consolidated view. Weekly status updates required 4 hours of manual work per PM.

Stack

Jira APIConfluence APIGoogle Apps ScriptBigPicture

Results

  • →Unified view of 50+ initiatives
  • →4h/week saved per PM
  • →Adoption by 12 teams in 2 months
GitHub →
03

CapLabour Capacity Planning

Apps ScriptGoogle SheetsPlanning

Capacity planning tool (v1 + v2) enabling managers to forecast workload and optimize resource allocation.

Problem

Capacity planning was done in Excel with outdated data. Over-allocations were detected too late, causing project delays.

Stack

Google Apps ScriptGoogle SheetsJira APICharts API

Results

  • →Capacity visibility across 8 teams (60+ people)
  • →40% reduction in allocation conflicts
  • →V2 delivered with rolling 12-week forecast
GitHub →
04

Kindly Reminder

Botn8nMonitoring

Compliance bot with automated monitoring. Smart reminders and real-time corrective action tracking.

Problem

Teams regularly forgot to complete compliance actions, leading to audit risks and time-consuming manual follow-ups.

Stack

n8nSlack APIGoogle SheetsCron scheduling

Results

  • →Compliance rate from 68% to 95%
  • →Zero manual follow-ups needed
  • →Real-time monitoring with escalation alerts
GitHub →
05

n8n Enterprise Workflows

n8nAutomationEnterprise

18 enterprise automation workflows covering onboarding, reporting, alerting, and cross-system data synchronization.

Problem

Dozens of repetitive manual processes between Slack, Jira, Google Workspace, and internal tools, generating errors and wasted time.

Stack

n8nREST APIsWebhooksPostgreSQLSlack API

Results

  • →18 workflows in production
  • →200+ hours/month saved
  • →99.5% uptime over 12 months
GitHub →
06

Jira Automation Rules

JiraAutomationCloud

18 Jira Cloud rules automating ticket management: assignment, transitions, notifications, and SLA tracking.

Problem

Manual Jira ticket management created status inconsistencies, forgotten assignments, and approximate SLA tracking.

Stack

Jira Cloud AutomationJQLJira REST APISmart Values

Results

  • →18 active rules across 5 Jira projects
  • →Triage time reduced by 70%
  • →SLA compliance improved from 78% to 96%
GitHub →
07

Automation Catalogue

DocumentationApps ScriptGovernance

Centralized registry of all organization automations with documentation, ownership, and performance metrics.

Problem

With 40+ automations in production, nobody knew exactly what existed, who was responsible, or if it was working correctly.

Stack

Google Apps ScriptGoogle Sheetsn8n APIJira API

Results

  • →40+ automations documented and catalogued
  • →Clear ownership for each workflow
  • →Health dashboard with outage alerts
GitHub →
08

IMOC RAG Incident Review Creator

n8nRAGAI

RAG-powered system that automatically generates structured incident post-mortems and vectorizes all JPROD incident data for instant retrieval.

Problem

Incident reviews required hours of manual work, resulting in inconsistent documentation, delayed learnings, and no searchable incident knowledge base.

Stack

n8nOpenAI APIPineconePostgreSQLSlack API

Results

  • →Incident reviews generated in <5 min vs 2+ hours manually
  • →All JPROD incidents automatically vectorized and searchable
  • →Tool in daily production use by the incident management team
09

Kallpa Running Coach

SaaSNext.jsAI

AI running coach SaaS generating personalized training plans using VDOT algorithm and Pfitzinger methods. Live at app.kallpa.run.

Problem

Runners had no affordable access to structured training plans adapted to their real level, race calendar, and weekly time constraints.

Stack

Next.js 14SupabaseVercelClaude APIStrava API

Results

  • →Full SaaS in production at app.kallpa.run
  • →4 locales (FR/EN/ES/PT), Strava & wearable integration
  • →VDOT + Pfitzinger algorithm for race-specific periodization
GitHub →
10

Jobhunter-AI

PythonAIMCP

Semi-autonomous job search pipeline with AI-powered multi-bloc scoring. Scans Gmail, job boards, and LinkedIn to surface high-match opportunities.

Problem

Job hunting is noisy and time-consuming. No tool matches opportunities against a nuanced personal profile across multiple dimensions.

Stack

PythonClaude APIMCPGmail APIPostgreSQL

Results

  • →Multi-profile support, Docker-ready, open-source
  • →AI scoring across 6 dimensions with auto-match ≥80
  • →Pipeline actively used in real job search with measurable signal
GitHub →

Have a similar project in mind?

Book a free AI audit