ATS Resume Builder
Personal Project
A production-ready resume editor with live ATS scoring, paid secure exports, and subscription billing — built and shipped end-to-end.
PythonFlaskJWTStripeRazorpaySQLitePWA
AI Engineering & Innovation Lab
Seven personal resume/hiring-tooling projects — solo-built, not enterprise deliverables — extending backend engineering into prompt engineering, RAG, and LangChain.
7 personal AI projects · Solo-built, not enterprise production
Personal AI projects
7
Applied in shipped work
2
Still exploring
6
No AI project below carries a verified date, so this is a conceptual progression — the same “widely-recognized shape, verified technology at each layer” treatment the Architecture Gallery already uses for its own Java-backend layering — rather than a dated career timeline.
Prompt Engineering
Retrieval-Augmented Generation (RAG)
LangChain
Every project below was designed, built, and shipped by one person — not an enterprise deliverable. What varies between them is how deeply each one is documented here; where that’s still thin, the card says so rather than filling the gap.
Showing 7 of 7 projects
A production-ready resume editor with live ATS scoring, paid secure exports, and subscription billing — built and shipped end-to-end.
A tool for extracting structured data (skills, experience, education) out of unstructured resume documents.
A system for scoring how well a resume matches a specific job description.
A tool that rewrites or scores resume content to better align with a target role.
A tool that generates a tailored cover letter from a resume and a target job description.
Retrieval-augmented generation work applying LangChain-based pipelines to ground LLM output in real source data.
A personal product built to make part of the hiring workflow easier — full scope to be confirmed.
Grouped by the same seven categories most AI-engineering stacks get organized by. Only technologies verified in one of the seven projects’ own stack appear as badges below — a category with nothing verified yet says so instead of being dropped, so the shape of what’s confirmed and what isn’t stays visible.
Prompt Engineering
Frameworks & Languages
Two maxims, each tied to one already-verified project — not six, because only two of the commonly-cited AI engineering principles are actually reflected in shipped work today.
Retrieval over memorization.
RAG Applications grounds LLM output in retrieved source content instead of relying on a model's parametric knowledge alone — the same idea the Architecture Gallery's own verified pattern for this describes in solution-shape terms.
Read the case study →Security even in a solo project.
ATS Resume Builder's paid tier runs JWT-based auth with OTP verification and signed, time-limited download links — the same discipline applied to a personal project's payment/export flow as to an enterprise authentication system, not relaxed because nobody's reviewing it.
Read the case study →Named but not included
What’s actually been applied in shipped, personal work, and what’s just an area of continued interest — kept as two distinct lists rather than one, so curiosity never reads as production experience.
Applied
Exploring — not yet verified in shipped work