Harshal Patil

Genora is the largest thing I have built, and it is the reason the tallest building in this city has my product’s name on it.

It is a premium, production-grade Next.js ecosystem containing eight advanced AI-powered tools, each of which would be a reasonable product on its own:

• GENxCode — a developer suite covering more than 40 languages.
• Chat with PDF — document intelligence over uploaded files.
• Social Media Studio — copy and asset generation for post workflows.
• SEO Generator — structured metadata and content briefs.
• Résumé Builder — structured document generation.
• Math Solver — step-by-step working, not just answers.

The part I am actually proud of

Not the tools — the state layer underneath them.

Eight tools that each stream results, hold session history and need to stay responsive while work is in flight is a state-management problem long before it is an AI problem. I built a custom Firestore state-management layer so that every tool gets real-time interactivity without each one reinventing its own subscription handling, optimistic updates and cleanup.

That is the work. The model calls are the easy part; keeping eight independently-streaming surfaces coherent in one app, on one auth session, with one billing context, is where the engineering went.

Stack decisions

Next.js 15 and React 19 for the app shell and server components. TypeScript throughout — with eight tools sharing one state layer, the types are the only thing stopping a refactor in one tool from silently breaking another. Firebase for auth and Firestore for state. TailwindCSS so the eight tools look like one product instead of eight.

It is fully responsive, because a developer tool that does not work on the phone in your hand at 11pm is a tool you will not reach for.

Cloud Yard is the district where the server stacks are labelled with the services they stand for, because this project is infrastructure rather than an interface.

An end-to-end, production-grade cloud deployment on AWS:

Why six services and not one

Any one of these is a tutorial. Wiring all six into a system that stays up is a different exercise, and it is the one worth talking about in an interview.

The interesting problems were never “how do I launch an EC2 instance.” They were the seams: getting CloudFront to cache what should be cached and pass through what must not be, getting ACM’s validation to actually complete against Route 53’s records, putting RDS somewhere that is reachable by the app and by nothing else, and deciding what belongs on S3 versus what belongs in the database.

What I took away from it

Infrastructure work is mostly about failure modes. The deployment that matters is not the one that works on the day you build it — it is the one that still works after a certificate expires, a region has a bad hour, or traffic arrives from somewhere you did not plan for.

That instinct is the direct reason the game layer of this portfolio has an eight-step performance bail-out ladder and a guaranteed HTML fallback underneath it. Same habit, different stack.

The Workshop is the one district with tools on the bench instead of a product in the window, because BugToBuild is not a product. It is the umbrella my own engineering work lives under — the workspace, the toolchain, and the habits.

Every other project in this city was built inside it. Including this city.

Why it has a name at all

Naming your workspace sounds like an affectation until you have four projects running at once and a scattered set of one-off scripts you rewrite every time.

Giving it a name forced a decision: anything I have solved twice gets promoted out of the project it started in and into the shared workspace, with a real interface and a real place to live. Anything I have solved once stays where it is. That single rule is most of what separates a toolchain from a folder of scripts.

What lives in here

Scaffolding and project bootstrapping, local automation for the repetitive parts of setup and deploy, and the shared TypeScript configuration and conventions that every project inherits so I am not re-deciding the same things monthly.

The name is the method

A bug report is the cheapest specification you will ever be handed. It is a reproducible statement about a gap between what the system does and what it should do — which is exactly the thing a feature request usually is not.

So the pipeline runs one way: bug, to reproduction, to the smallest change that closes it, to a build. Not “bug to patch.” Bug to build — because if closing it did not leave the thing shippable, it is not closed.

The Archive is a records building: long shelves, indexed drawers, one reading desk. That is the honest shape of this project.

Ancestriq turns family records into personalised ancestry narratives — you give it what is known about a lineage, and it produces a readable story rather than a table. The interesting part was never the generation. It was the records.

The actual problem

A generative model handed loose family records will confidently produce a plausible narrative that is not in the records. For anything genealogical, a plausible answer is worse than no answer, because an invented ancestor reads exactly like a real one — and this is someone’s family.

So the work went into the layer underneath: getting records into a structure where every claim in the output traces back to a specific input, and where the model’s job is narrowed to phrasing what it was given rather than deciding what is true. If a detail is not in the records, the story does not get to have it.

What I learned that carried forward

Two things, and both of them are visible elsewhere in this city.

The first is that the record layer is the product. Swapping the model is an afternoon. Getting the data into a shape where the output is constrained by it is the project. I have not seen a GenAI application where this was not true.

The second is knowing when not to reach for a model. That lesson is the direct reason this portfolio contains no AI at all — no chatbot, no assistant, no inference at runtime. Every question a recruiter might ask is answered by authored text in the Dossier, and every behaviour in the game is deterministic logic. Having built with these models, I have a clear view of where they earn their latency and their bill. A portfolio is not one of those places.

Full Stack Developer with hands-on experience building and deploying web applications, AI-integrated tools, and cloud infrastructure.

Proficient in JavaScript (React, Node.js, Express), Python, and Dart/Flutter with production deployments on AWS and Firebase. Adept at end-to-end product development — from UI/UX implementation to backend APIs and CI/CD workflows.

Based in Chhatrapati Sambhajinagar, Maharashtra, India. Available for collaboration — freelance and contract work.