Learn to finish the whole loop
Small automation, API, and data projects taught me to carry an idea from input to a result someone could actually use.
ATISHAY
The systems engineer.That's Atishay.
02 Featured
systems
systems
11 Public
builds
builds
Atishay Jain / Computer Engineering Atishay Jain / AI + product systems
AI Systems,
Applied
Differently.
● Practical◆ Reliable△ Systems-minded✦ Product-aware■ Evidence-led
I turn models, APIs, state, and interfaces into complete workflows with explicit fallbacks and honest limits.
Proof anchors Scroll to follow the build pathProjects (&)
How I'm Improving
Each project added a harder problem: first finishing the tool, then connecting the system, then making its behavior reliable.
Move beyond isolated screens
The Loop and Hybrid forced me to connect interfaces with persisted state, collaboration, routing, and explicit fallback paths.
Make AI behavior deliberate
RAG, resume tailoring, and Mahoraga moved me from calling models to designing retrieval, routing, memory, reward signals, and evaluation.
Harden what ships
Braillix, PlantBrain, and ReceiptSplit explore accessible learning, cited operational knowledge, and shared-expense coordination with explicit evidence boundaries.
Selected work Built to perform,
Built to perform,
made to be inspected.
Three focused systems. Start with each project's UI, then inspect the decisions and evidence behind its build.
Browse the project gallery
01/ 03
Proof on file: Lesson-to-Braille software flow PRODUCT UI · SIMULATED DISPLAY FEATURED / COLLABORATIVE PROTOTYPE
Braillix
Math teachers need a classroom flow that carries a changing lesson from the board to a student's refreshable Braille display.
What I learnedAccessible workflows must keep the written lesson, spoken readback, and selected Braille frame in sync.
Proof on file: Cited evidence workflow PRODUCT UI · SYNTHETIC RECORDS FEATURED / SYNTHETIC-DATA PROTOTYPE
PlantBrain
Operational troubleshooting needs answers that make their supporting evidence easy to inspect.
What I learnedGrounded AI is a product workflow: retrieval, evidence, stale states, and explicit unknowns all matter.
Proof on file: Room-to-settlement workflow PRODUCT UI · SAMPLE ROOM DATA FEATURED / PRODUCT SYSTEM
ReceiptSplit
Groups need to move from a shared receipt to an agreed split without treating a payment hand-off as proof.
What I learnedMoney correctness is a product contract: allocate paise deterministically and keep settlement claims explicit.
What
You Get?
and combined—turninguncertaintyintoausefulproductexperiencethatfeels
01 Foundations
Learned to reason from algorithms, data shape, and query structure before reaching for frameworks.
C++/DSA / ML basics / SQL foundations
02 Product Engineering
Learned that useful products connect discovery to persisted coordination, not isolated screens.
The Loop
03 AI Systems
Learned to route ambiguity through rules, local inference, memory, and fallback paths.
Hybrid GenAI Transaction Categorizer
04 Adaptive Systems
Learned that environments, reward signals, and evaluation limits shape agent behavior.
Mahoraga
05 Current Direction
Applying those lessons to accessible interaction, cited evidence, and explicit money-state boundaries.
Braillix / PlantBrain / ReceiptSplit
Working set Tools
Tools
That Deliver
The stack changes with the system. These are tools traced back to real project work—not a wall of self-rating badges.
01
Languages
05 tools PythonTypeScriptJavaScriptC++SQL
02
Frontend
06 tools ReactNext.jsAstroViteTailwind CSSshadcn/ui
03
Backend / APIs
05 tools FastAPINode.jsREST APIsWebSocketsUvicorn
04
AI / ML
07 tools ONNX RuntimeDistilBERTQwen2.5 GGUFllama.cppPyTorchGymnasiumLoRA
05
Data / Databases
04 tools PostgreSQLSQLiteSupabaseSQLAlchemy
06
DevOps / Deployment
04 tools DockerGitHub ActionsGitHubRender
07
Product Systems
06 tools Auth/JWTRealtimeOCR pipelinesUPI deep linksMapsRecommendations
Capabilities are shown through the projects that made them real—not detached keyword claims.
AI systems architecture
PlantBrain / Mahoraga / Hybrid Categorizer / AI Resume Tailoring Engine
Product engineering
Braillix / PlantBrain / ReceiptSplit / The Loop
Automation
Resume Engine / Webcam Motion Alert / Music Web Scraper
Data and ML foundations
Minimal RAG / Global Super Store / Fast and Curious / Linear Regression
Full-stack implementation
Braillix / PlantBrain / ReceiptSplit / The Loop / Hybrid Categorizer
C++/DSA foundations
Coursework and resume signal
Project gallery
More builds, focused constraints
Earlier featured case studies sit alongside supporting projects, with the same evidence-first framing.
Proof on file: Current demo UI and reward loop GALLERY / PROTOTYPE
Mahoraga
Adaptive RL/game-system prototype with reward shaping, curriculum pressure, and LLM integration.
An adaptive-combat prototype shaped by rewards; its public UI shows a disconnected sample encounter.
Proof on file: Routing and memory flow GALLERY / LOCAL AI SYSTEM
Hybrid GenAI Transaction Categorizer
Privacy-first transaction categorizer using merchant overrides, ONNX DistilBERT, and offline Qwen fallback.
Categorize transactions without making the LLM the default path.
Proof on file: Workflow and system diagrams GALLERY / PRODUCT SYSTEM
The Loop
Campus event platform for discovery, RSVP, friends, chat, maps, and carpool workflows.
Connect event discovery, interest context, RSVP, social coordination, chat, and carpool planning.
SecondaryRepo and generated SDE resume verified
Proof on file: Generated public resume SECONDARY / REPO AND GENERATED SDE RESUME VERIFIED
AI Resume Tailoring Engine
Privacy-first, deterministic resume and cover-letter generation.
Tailor application documents against a job description without inventing profile claims or sending personal data to an external AI API.
SecondaryRepo verified, compact support
Proof on file: Supporting workflow artifact SECONDARY / REPO VERIFIED, COMPACT SUPPORT
Webcam Motion Alert System
OpenCV and alerting workflow for motion-triggered notifications.
Detect motion and trigger useful alerts with cleanup instead of a passive camera script.
ATISHAY ®
Contact Got a system
Got a system
worth building?
For software, AI systems, or product engineering opportunities, inspect the work and use the direct path that fits.