> boot maharshi.profile
> load identity: AI + full-stack engineer
> load education: M.S. Computer and Information Science @ Texas A&M University-Corpus Christi
> load pattern: rough idea -> prototype -> evaluation -> usable product
> load output: software that can be demoed, shipped, explained, and handed off
> status: buildingI am Maharshi Barot. I build at the point where AI stops being a prompt and becomes a working product: a mobile interface, a backend API, a database model, a route-aware map, realtime chat, payments, liveness checks, dashboards, or a workflow someone can actually use.
This profile is intentionally not a normal badge wall. It is a small map of how I think, build, test, explain, and hand off systems.
| Runtime module | Loaded capability | Real-world output |
|---|---|---|
idea.kernel |
Converts unclear asks into product shape | scoped flows, requirements, demos |
ai.engine |
OpenAI APIs, prompt design, RAG thinking, LLM evaluation | assistants, tutors, automation loops |
product.shell |
React Native, React, Next.js, TypeScript | mobile apps, dashboards, interfaces |
system.drivers |
Firebase, FastAPI, Django, Node.js, .NET, SQL | auth, APIs, data, realtime features |
reality.adapters |
Mapbox, Stripe, OpenCV, Docker, GitHub Actions | maps, payments, liveness, CI/CD |
maharshi build --mobile React Native + Firebase + maps + payments
maharshi build --ai OpenAI APIs + prompt loops + evaluation
maharshi build --backend FastAPI / Django / Node / .NET + PostgreSQL
maharshi build --vision OpenCV + embeddings + liveness + analytics
maharshi explain --handoff docs that let the next person continue the system|
01. Intelligence Layer OpenAI APIs, prompt engineering, RAG thinking, LLM evaluation, NLP, workflow automation. |
02. Product Layer React Native, React, Next.js, TypeScript, Tailwind CSS, Bootstrap, usable frontend flows. |
03. Systems Layer Firebase, Node.js, Express, Django, FastAPI, .NET 8, REST APIs, auth, RBAC, CI/CD. |
04. Reality Layer PostgreSQL, MongoDB, Firestore, SQLite, Mapbox, Stripe, OpenCV, Docker, GitHub Actions. |
flowchart LR
A["Messy idea"] --> B["Product shape"]
B --> C["Data model"]
C --> D["API contract"]
D --> E["Interface"]
E --> F["AI or automation loop"]
F --> G["Evaluation"]
G --> H["Demo-ready system"]
H --> I["Documentation and handoff"]
R["React Native / React"] --> E
P["Python / Node / .NET"] --> D
L["OpenAI APIs / RAG / NLP"] --> F
DB["Firebase / PostgreSQL / MongoDB"] --> C
| Mission | What I built | System texture |
|---|---|---|
| AI Tutor Mobile App | A mobile AI tutoring experience powered by LLM APIs | React Native, TypeScript, OpenAI APIs, prompt iteration |
| RamayanaGPT | A domain-focused conversational assistant | LLM APIs, answer quality, prompt design, user-facing AI behavior |
| Campus Ride Pooling Mobile App | Campus ride creation, chat, auth, payments, identity, and route-aware matching | React Native, Firebase, Node.js, Mapbox, Stripe |
| Face Recognition Attendance System | Attendance platform with face embeddings and liveness detection | Python, FastAPI, OpenCV, JWT auth, RBAC, analytics, CSV/XLSX export |
| AI-Powered Retail Investor Dashboard | Market-data dashboard with realtime insights and interactive visuals | Python, React, SQL, REST APIs, data aggregation |
|
ai-tutor-app TypeScript-first mobile AI work. Useful if you want to see how I shape AI into a product surface. |
face-attendance-app Applied computer vision, backend auth, liveness checks, dashboards, and export workflows. |
|
KAN-ODEs Research-oriented machine learning and numerical experimentation. |
reinforcement-learning Reinforcement learning experiments and applied AI practice. |
| Strand | Evidence from my work | Direction I keep pushing |
|---|---|---|
| Product instinct | AI Tutor, Campus Ride, dashboards | make technical systems feel usable |
| AI depth | LLM apps, RAG thinking, evaluation loops | make AI outputs more reliable |
| Backend discipline | REST APIs, auth, RBAC, exports | make frontends easy to integrate |
| Realtime thinking | Firebase, chat, ride creation, live data | make systems respond like products |
| Applied ML | face embeddings, liveness checks, RL, KAN/ODE experiments | connect research ideas to working tools |
| Signal | Tools I use | What I use them for |
|---|---|---|
| AI / LLM | OpenAI API, prompt engineering, RAG, LLM evaluation, NLP | assistants, tutoring flows, automation loops, quality tuning |
| Mobile | React Native, Firebase, Mapbox, Stripe | cross-platform product flows, realtime features, maps, payments |
| Frontend | React, Next.js, TypeScript, Tailwind CSS, Bootstrap | dashboards, app screens, interactive visual systems |
| Backend | Node.js, Express, Django, FastAPI, .NET 8 | REST APIs, auth, service integration, data pipelines |
| Data | PostgreSQL, MongoDB, Firestore, SQLite, SQL | product data models, analytics, storage, search, exports |
| Vision | Python, OpenCV, embeddings, liveness checks | face recognition, attendance systems, verification workflows |
| Dev Tools | Docker, Git, GitHub Actions, JWT, CI/CD | deployment readiness, automation, reproducible engineering |
- M.S. in Computer and Information Science at Texas A&M University-Corpus Christi
- B.Tech in Computer Science and Engineering from Pandit Deendayal Energy University
- Diploma in Computer Engineering from Government Polytechnic, Ahmedabad
- turning an AI demo into an actual product flow
- connecting mobile apps to Firebase, APIs, maps, payments, and realtime data
- building backend services that frontend teams can understand and use
- explaining technical systems to non-technical users without losing the important details
- documenting the build so the next person is not trapped inside my head
Open the hidden build notes
Design rule:
A project is not finished when the code runs.
It is finished when the user understands it, the next engineer can continue it,
and the system can survive a real demo.
Personal constraint:
Make the work memorable without making it confusing.Open to AI product engineering, full-stack development, mobile apps, backend APIs, and applied ML projects.