Learning Journey | Robert Consulting

🎬 Plex Movie Recommendations

AI-Powered Movie Recommendation System

✅ Completed 🌐 Live Demo

🎯

95%

Accuracy

⚡

2.3s

Response Time

🤖

Algorithm

📊

1000+

Movies

Project Overview

Built an intelligent movie recommendation system using machine learning algorithms and modern web technologies. The system learns from user preferences and viewing history to provide increasingly accurate movie suggestions.

🧠

Machine Learning

Advanced recommendation algorithms that adapt to user preferences

⚡

Real-Time Processing

Responsive system that updates recommendations instantly

🎨

Modern UI/UX

Clean, intuitive interface with smooth animations

Live Demo

Experience the AI-powered movie recommendation system in action:

🎬 Explore Plex Recommendations

Loop

Interview Observability Platform

A personal feedback pipeline: secure-by-default, automated, measurable. DevSecOps-on-yourself—treat interview content like production data and improve with real metrics.

Principles

Data minimization

Store what you need; redact the rest. No raw PII in analytics.

Security by default

Encryption everywhere, least privilege, audit logs, retention.

Repeatable + portable

Terraform + CI/CD + versioned configs. No snowflake scripts.

Measurable outcomes

Dashboards and trend lines, not vibes. Same as production.

What it is

A self-improvement feedback pipeline: capture interview-related content, transcribe and normalize it, redact PII, run analysis for structured metrics and coaching output, then store and index results. The pipeline runs automatically after each session and feeds into dashboards and trend lines—same operational mindset as production systems.

Resources used

Built on AWS with a DevSecOps/SRE lens:

Storage: Object storage for raw and processed artifacts; structured store for metadata and scores; optional full-text search.
Processing: Managed transcription; PII detection and redaction; orchestrated analysis jobs (serverless or container).
Observability: Metrics, dashboards, and alerts for pipeline health, latency, and cost.
IaC and CI/CD: Terraform for provisioning; policy-as-code and security scanning in the pipeline; versioned configs and drift detection.

Security & reliability (DevSecOps/SRE)

Designed so the story is interview-worthy:

Identity: Least-privilege roles per stage (ingest, process, redact, analyze, read-only).
Encryption: At-rest and in-transit; dedicated keys where appropriate; secrets in a managed store.
Network: Controlled egress; service endpoints where possible; allowlist-based access.
Governance: Retention and lifecycle rules; audit logging; optional lock/immutability.
CI/CD guardrails: Plan → policy checks → apply with approval; nightly drift detection; security and compliance scanning on infra code.

Interview content is treated like sensitive data: minimized, redacted before analysis, and protected with the same controls you’d use in production.

What you can say in interviews

“I built a secure, automated interview feedback pipeline: capture, transcribe, PII redaction, and step-orchestrated analysis, deployed with Terraform and policy-as-code in CI/CD. It produces measurable communication metrics and coaching artifacts so I improve continuously—the same way I run production systems.” DevSecOps and SRE discipline applied to self-improvement.

🏗️ Infrastructure as Code

Real infrastructure managed with Terraform

4,758

Lines of Terraform Code

Infrastructure as Code

129

AWS Resources Managed

All managed by Terraform

AWS Services Integrated

Multi-service architecture

Security Resources

WAF, IAM, Encryption

🚀 Infrastructure Features

🏢 Multi-Account Architecture

AWS Organizations with separate client accounts and cross-account access

🔒 Security & Compliance

WAF protection, IAM roles, encryption, and security monitoring

🗄️ State Management

Remote state storage with S3 backend for centralized state management

📚 Weekly Learning Reflection

💡 Top Lessons Learned This Week

🎯 Focus Areas Next Week

📊 Key Metrics This Week

📦 Separate Content & Infrastructure Versioning

📝 Content Versioning (CalVer)

Uses timestamp-based versioning (YYYYMMDDHHMMSS) for frequent content updates. Each deployment gets a unique version number that reflects when it was created.

Format: v20251120113439
Updated on every content deployment
Stored in VERSION and website/version.json
Enables rollback to specific content versions

🏗️ Infrastructure Versioning (SemVer)

Uses semantic versioning (MAJOR.MINOR.PATCH) for infrastructure changes. Only incremented when infrastructure code or configuration changes.

Format: 1.1.3
MAJOR: Breaking changes
MINOR: New features, backward compatible
PATCH: Bug fixes, backward compatible
Stored alongside content version in same files

💡 Benefits of Dual Versioning

Independent Release Cycles: Content can be updated frequently without triggering infrastructure deployments
Clear Change Tracking: Easy to identify what changed and when across separate repositories
Rollback Capability: Can rollback content or infrastructure independently
Better Coordination: Infrastructure team and content team can work independently with clear version boundaries

Continuous Learning Journey