Independent builder

Useful ideas.
Working software.

I’m Madhav Kadiyala. I build tools for creative work, connected teams, and everyday workflows.

Explore six projects Meet the builder

01 / EXPLOREAn idea

02 / BUILDMK + Codex

03 / REFINETest & iterate

A real task. A working system. Another iteration.

A closer look

Selected projects.

Interfaces you can inspect.
Engineering you can follow.

01 / Media & creative toolsLocal application

ClipForge

One long video.
A stack of possibilities.

A video clipping pipeline that connects transcription, moment selection, vertical reframing, and animated captions.

Python
FastAPI
faster-whisper
OpenCV
FFmpeg

ClipForge upload interface with a video URL field, upload button, clipping options, and Make clips action — Rendered from the actual local interface. Personal job history omitted.

Inside the build

The workflow

Ingest a video file or URL.
Transcribe and rank candidate moments.
Reframe to 9:16 and animate captions.
Export MP4, SRT, XML, or EDL.

The decisions

Connect the media stages into one usable local workflow. Optional model integrations strengthen ranking, while a heuristic fallback keeps the core pipeline usable without them.

Brand presets, thumbnails, audio enhancement, optional B-roll, voice-over, and YouTube scheduling extend the pipeline.

Build notes

My role: independent builder, working with Codex.

An end-to-end development sample produced a 1080 × 1920 clip with burned-in captions. External publishing integrations depend on configuration.

02 / Job-search systemsLocal runtime verified

JobClaw

A connected workspace
for the job search.

A job-search workspace that brings discovered roles, profile matching, tailored resumes, and application tracking into one dashboard.

Python
FastAPI
Jinja2
SQLite / PostgreSQL
Claude API

JobClaw dashboard running locally with overview cards, an empty matches list, and an inactive plan — Actual local dashboard with an isolated empty database. External services disabled.

Inside the build

The workflow

Collect roles, match them against a profile, and prepare tailored resumes. The workspace tracks applications, resume generation, and plan usage.

The architecture

FastAPI serves Jinja templates. Background schedules coordinate collection and generation; Claude integrations support matching and tailoring. SQLite supports local runs, with PostgreSQL configured for hosted use.

Private repository · access required

Verified locally

Seven existing product tests passed, covering routes, dashboard rendering, authentication fixtures, and source controls. The screenshot comes from the running app.

Live collection, Claude calls, resume generation, and production deployment were outside this verification.

03 / Trading research & automationLocal research system

Webull AI
Trading Bot

Every signal.
An inspectable decision.

A Python toolkit that separates trade ideas, risk checks, broker adapters, and reporting—so research decisions leave a clear record.

Python
pandas
NumPy
Webull SDK
LLM APIs

Open sample report

Actual trading report renderer showing three rejected synthetic test signals: low confidence, zero quantity, and unavailable market data — Actual report renderer with synthetic test fixtures. No trades placed or account data used.

Inside the trading system

Research before execution

Paper and replay modes use synthetic data by default. Opening-range breakout backtests accept minute-bar CSV data or an optional Databento historical feed with a download cost cap.

Strategy research includes multiple seeds, slippage stress tests, and simulated market regimes.

Independent risk checks

The risk engine checks confidence, sizing, eligible symbols, trading windows, and configured loss limits. An optional LLM can produce ideas; a deterministic fallback is also implemented.

The Webull adapter has order previews and a separately enabled execution path.

What this sample verifies

Three offline inputs were rejected by the actual risk engine and passed to its HTML report renderer. These component checks cover low confidence, invalid quantity, and missing data.

They are not an end-to-end trading test or evidence of live performance or profitability.

More working software

Systems in practice.

Three local applications.
Real workflows, synthetic data.

04 / Collaboration

TeamFlow

A shared task board with live updates, drag-and-drop moves, comments, and separate editor and viewer permissions.

ASP.NET Core
React / TypeScript
SignalR
SQLite

Build & verification

React Query refreshes board data through SignalR. EF Core persists tasks and comments; version checks handle conflicting moves.

3 xUnit and 2 browser tests passed locally, including two-session synchronization, permissions, duplicate submissions, and persistence.

Single-process local demo with synthetic accounts.

Private repository · access required

05 / Document workflows

DocuLens

Upload text, PDF, or DOCX files. Follow processing, search your library, and trace extractive previews back to their source.

ASP.NET Core
React / TypeScript
PdfPig
SQLite

Build & verification

A durable queue records processing events, bounded retries, and restart recovery. Search and downloads stay scoped to the document owner.

5 xUnit and 1 browser test passed locally, covering extraction, source citations, owner isolation, failure recovery, and persistence.

Summaries are cited excerpts. No generative model or cloud AI service is used.

Private repository · access required

06 / Order management

OrderDesk

Reserve stock, follow an order’s progress, and handle a payment retry through an explicit, inspectable timeline.

ASP.NET Core
React / TypeScript
EF Core
SQLite

Build & verification

Transactional reservations and repeat-safe commands protect stock. Owner and operator permissions control payment, cancellation, and shipping actions.

5 xUnit and 1 browser test passed locally, including concurrent reservations, oversell prevention, duplicate commands, and payment retry after restart.

Payments and fulfillment are simulated. One local process owns the workflow.

Private repository · access required

All three run locally with ASP.NET Core 10, React, and SQLite. Hosted CI and Azure deployment have not been verified.

Behind the work

Curiosity.
Then code.

I’m Madhav Kadiyala, an independent builder exploring what useful software can do.

My projects start with a task worth making easier. I work with Codex to turn an idea into an interface, connect the pieces, and keep improving how the system works.

I’m interested in the decisions underneath the screen: where work happens, how failures are handled, and what makes a workflow understandable.

Madhav Kadiyala MK.

Useful ideas.Working software.

Selected projects.

ClipForge

The workflow

The decisions

Build notes

JobClaw

The workflow

The architecture

Verified locally

Webull AI Trading Bot

Research before execution

Independent risk checks

What this sample verifies

Systems in practice.

TeamFlow

DocuLens

OrderDesk

Curiosity. Then code.

Useful ideas.
Working software.

Webull AI
Trading Bot

Curiosity.
Then code.