Experience

Internship

• Built a CTR Opportunity Scoring capstone: a decision-support model that ranks pages for SEO metadata review.
• Used DuckDB to query the ~78.8M-row Hugging Face warehouse (Parquet) without loading it into RAM, verifying grain and scale; modelled on a 30k-row anonymized starter slice.
• Compared a hand-written baseline, Logistic Regression, and Random Forest on a client-holdout split: Random Forest reached Precision@50 of 0.640 vs 0.260 for the rule.
• Demonstrated client-overlap leakage (naive split 0.940 vs honest 0.640) and planted two deliberate leaks (trend_pct, ctr) to check the validation harness.
• Turned scores into six reason-coded action archetypes with human-review requirements and an explicit list of what should never be automated.
• Wrote and published the research paper, and completed Anthropic Academy courses on AI fluency and Claude API usage.

Apprenticeship

• Worked through ML mathematics, vector computation, classical ML, and deep-learning fundamentals with hands-on projects.
• Built and served ML applications with FastAPI and containerized them with Docker.
• Completed and defended the final project in July 2026.

Mandatory academic internship

• Independently designed and implemented the architecture of the AcademiaOS MVP.
• Built backend services with FastAPI and Celery and a Next.js frontend.
• Implemented HTTP-only cookie authentication and role-based access control (RBAC); used Docker in development and deployment.

Contract

• Shipped features, bug fixes, and a UI revamp, and supported the Google Play Store launch of the Academia mobile app.
• Maintained existing Flutter codebases and migrated corporate websites to Next.js.

Contract

• Built an e-commerce admin panel with React, MUI, and Redux-Saga.
• Developed Next.js frontends integrated with existing PHP backends.

Internship

• Built Flutter apps with Firebase Authentication, local persistence, and BLoC state management.