Boggavarapu Yuva Satya Kunaal — AI Engineer, Multi-Agent Systems Builder & SaaS Founder
A B.E. AI & Data Science graduate from CBIT, founder of SQLumina, and builder of production-grade systems across AI pipelines, data engineering, automation, and SaaS.
Systems that think. Products that ship.
Boggavarapu Yuva Satya Kunaal is an AI & Data Science engineer who doesn't prototype and move on — he designs, builds, ships, and owns end-to-end systems. From multi-agent AI pipelines to production-grade SaaS with real users, the work spans the whole reel: idea → architecture → launch.
The Features
SQLumina — Large-Scale SaaS (2026)
A full structured SQL learning ecosystem — schema-aware lessons, practice environments, progressive challenges, and intelligent guidance. Built from the ground up to be the definitive place to master SQL. Production-grade SaaS architecture. Deployed; launch incoming.
Tech: SaaS, Full-Stack, SQL Learning, Production Architecture.
AI Interview SDK — Open-Source SDK (2026)
An open-source (MIT) TypeScript/React SDK that drops an AI-scored interview into any app as a single React component — your own AI keys, backend, and database, so the maintainers never see or store candidate data. A pnpm + Turborepo monorepo of 10 published npm packages: a framework-agnostic core with a dynamic follow-up engine, HMAC-signed server-mode scoring, five provider adapters (OpenAI · Claude · Gemini · Deepgram · ElevenLabs) with automatic failover, and a CLI with an Interview Simulator + Bias testing harness.
Tech: TypeScript, React, Node.js, Turborepo, Vitest.
AI Interview SDK live demoRepository File Count — Open-Source Tool (2026)
A fast, no-clone way to see what's actually inside a GitHub repository. Paste a repo URL, sign in with a GitHub token, and get a full file-tree breakdown in seconds — total files, real code vs. docs/assets/config, and a per-extension count — without cloning it or setting anything up locally. Walks the repo tree straight from the GitHub API, entirely in the browser; the token never touches a server.
Tech: TanStack Start, React, TypeScript, Tailwind CSS, GitHub API.
Repository File Count live demoHabit Ink — Minimal SaaS (2026)
A minimal, intentional habit-tracking SaaS for people who want clarity over clutter. No noise, no gamification gimmicks — just a clean daily rhythm that actually sticks. Live and growing.
Tech: React, Firebase, Tailwind, Framer Motion.
Habit Ink live demoAI Flowchart Studio — Multi-Agent AI (2026)
Plain English to structured diagrams via a 4-stage Multi-Agent Orchestration pipeline: Orchestrator → Logic Parser → Generator → Syntax Validator. BYOK privacy architecture, SSE real-time streaming, 50-step undo/redo, export to PNG / SVG / Mermaid.
Tech: FastAPI, Gemini, Mermaid.js, Vanilla JS, Render.
AI Flowchart Studio live demoCommitAI — Developer Tooling (2026)
A local-first AI Git assistant. Reads staged diffs, prompts Ollama, normalizes output into Conventional Commits, writes the changelog, and completes the workflow — entirely on your machine. Built for the dev.to Gemma 4 Challenge (completion badge).
Tech: Python, Ollama, Gemma 4, Rich, Git Hooks.
CommitAI source codeKunaal's Illustrations — Open-Source Art (2026)
An open-source Agent Skill guiding AI agents to generate clean, hand-drawn in-text illustrations — defining character IPs, style DNA, composition patterns, and QA rules. A complete visual language, not a prompt.
Tech: Agent Skill, Gemini, Antigravity.
Kunaal's Illustrations source codeQuery Forge AI — NL → SQL Engine (2025)
A production natural-language-to-SQL engine with schema-aware prompting. Understands database structure to generate accurate, optimized queries from plain English.
Tech: FastAPI, MySQL, Mistral, Python.
Query Forge AI source codeAnalyzeQuestion — RAG System (2025)
An AI-driven coding pattern detection system built with a retrieval-augmented pipeline for grounded analysis.
Tech: RAG, LLM, Python.
AnalyzeQuestion source codeERP Intelligence Engine — Data Engineering (2025)
An end-to-end ETL pipeline with OCR-driven extraction automating the most time-consuming academic data workflows — saving 8+ hours every week.
Tech: Python, OCR, ETL, Automation.
ERP Intelligence Engine source codeChatWithPDF — PDF Q&A (2025)
A PDF question-answering system with OCR grounding and LLM accuracy for reliable document interaction.
Tech: RAG, OCR, LLM.
ChatWithPDF source codeAI Voice Desktop Assistant — Voice Agent (2025)
A voice-controlled desktop assistant for macOS — say "open terminal" or "close downloads folder" and it converts speech to text, parses intent with an LLM, fuzzy-matches the right app, file, or folder (typo-tolerant), and executes the action via AppleScript.
Tech: Python, Mistral (OpenRouter), Speech Recognition, AppleScript.
AI Voice Desktop Assistant source codeAutomation Systems — Workflow Automation (2025)
Certificate generation, attendance monitoring, and HR workflow automation systems built around practical operational needs.
Tech: Make.com, Google Sheets API, Gmail API, Google Drive API.
Automation Systems source codeAutomation Systems second source code
The Craft
The Shooting Schedule
From the opening frame to the feature — school, CBIT, and the work that followed.SCROLL TO ADVANCE →
10th Class — 10 / 10 CGPA · Sri Chaitanya Techno School
Intermediate, MPC — 970 / 1000 · Sri Chaitanya Junior Kalasala
Began B.E. AI & Data Science · CBIT
First build wave — NL→SQL & RAG engines, OCR pipelines, a voice agent, and automation systems
AI Interview SDK (open source · 10 npm packages) · Repository File Count · AI Flowchart Studio · CommitAI · Habit Ink
First peer-reviewed publication — "GodForge" — Springer, Lecture Notes in Networks and Systems (ICTCS 2025)
Graduated CBIT — 9.16 / 10 · First Class with Distinction
SQLumina deployed — large-scale SQL learning SaaS, launch incoming
Laurels & Honors
Published Work
Beyond the shipped work — peer-reviewed research, in print.