class DevParikh:
def __init__(self):
self.name = "Dev Parikh"
self.degree = "B.Tech Computer Engineering @ CHARUSAT University (2022–26)"
self.cgpa = "9.61 / 10"
self.location = "Gujarat, India"
self.email = "devparikh200479@gmail.com"
@property
def highlights(self):
return [
"2x GATE Qualified — DS&AI (AIR 6,505) + CSE (AIR 17,015) [2025]",
"2x Springer-Published Research Author — ICTCS 2024 / LNNS Series",
"Open Source Contributor — Accomplish AI (MIT) · +895/-2 lines merged to main",
"ML Club Coordinator @ CHARUSAT University",
"Co-founder @ Lekh.ai — Live AI Document Generation Platform",
"Software Engineer Intern — AI chatbot serving 100+ UK employers",
"5x NPTEL Elite Certified (Top 5% Nationwide)",
"Kaggle Competitor — Hull Tactical · March ML Mania · AI Math Olympiad",
"Top 75 / 2,200+ Teams — Odoo x CHARUSAT Hackathon",
]
|
B.Tech — Computer Engineering 🏛️ CHARUSAT University, Gujarat, India 📅 2022 – 2026 🎯 CGPA: 9.65 / 10 🔬 ML Club Coordinator 🏅 UGSF Research Fellow (Funded) |
GATE 2025 — Dual Qualification
|
💬 Languages
🤖 AI / ML / NLP
🌐 Web & Backend
🗄️ Databases & Infrastructure
Co-founder & Full-Stack Developer
The all-in-one AI document platform for elite academic applications — SOPs, Resumes, LORs, Research Papers, and IEEE SRS specs — built for students targeting top global universities.
| Engine | Capability |
|---|---|
| 🎯 SOP Engine | University-personalized Statement of Purpose |
| 📄 Precision Resume | Ivy League-standard academic CV |
| ✉️ LOR Engine | Faculty-grade Letters of Recommendation |
| 📐 SRS Generator | IEEE 830-1998 compliant Software Requirements Spec |
| 🔬 Research Paper | Structured academic papers with full methodology |
| 📚 Citation Engine | APA · MLA · Chicago · Vancouver auto-formatting |
Stack: Next.js TypeScript OpenAI API Anthropic API Multi-step LLM pipelines DOCX/PDF export
Full-Stack AI/ML Developer
Dink. Drive. Dominate. — Real-time computer vision analytics for competitive pickleball.
| Feature | Detail |
|---|---|
| 👁️ YOLO Detection | Real-time player & ball tracking |
| 📐 Kalman Filtering | Smooth multi-frame trajectory prediction |
| 📊 Analytics Dashboard | Shot heatmaps · rally stats · movement patterns |
| 🗄️ Data Layer | PostgreSQL for match history & player profiles |
| ⚛️ Frontend | React-powered responsive UI |
Stack: Python YOLO Kalman Filters React PostgreSQL
NLP Developer
- University rank predictor — 90% accuracy, used by 100+ students
- Full RAG pipeline:
Flask·Pinecone·Cohere·MongoDB
Active competitor across data science and ML challenges. Visit the profile for live rankings.
| Competition | Domain |
|---|---|
| 🏦 Hull Tactical — Market Prediction | Finance · Time Series · Algorithmic Trading |
| 🏀 March Machine Learning Mania | Sports Analytics · Bracket Prediction |
| 🧮 AI Mathematical Olympiad — Progress Prize | LLM Reasoning · Mathematical Problem Solving |
Peer-reviewed papers in Springer Nature — Lecture Notes in Networks and Systems (LNNS), presented at ICTCS 2024 (International Conference on Information and Communication Technology for Competitive Strategies).
Dev Parikh, Nikita Bhatt, Dhaval Bhoi
📘 Springer LNNS Vol. 1323 · pp. 307–316 · ICTCS 2024 🗓️ First Online: 14 August 2025 · 👁️ 125 Accesses
In-depth analysis of OCR techniques for extracting and distinguishing Gujarati and English text from bilingual documents. Proposes advanced preprocessing, text segmentation, and language identification. Addresses Gujarati script's complex structure, diacritical marks, and conjunct letters — achieving 90.75% (English) and 91.16% (Gujarati) extraction accuracy.
Jay Parmar, Dev Parikh, Amit Thakkar, Dhaval Bhoi
📘 Springer LNNS Vol. 1324 · pp. 225–235 · ICTCS 2024 🗓️ First Online: 01 October 2025 · 👁️ 101 Accesses
Comparative study applying NLP techniques to email spam classification. Benchmarks FastText against Bag-of-Words + Random Forest using tokenization, stopword removal, and stemming for preprocessing. Demonstrates significant improvement in spam filter resilience and scalability for real-world email security.
🖥️ Accomplish AI — Open Source AI Desktop Coworker
Contributor · TypeScript · Chrome DevTools Protocol (CDP) · Electron
Accomplish is an MIT-licensed open-source AI desktop agent (Electron + React) that automates file management, browser tasks, and document workflows locally.
PR #480 → Merged into main via PR #769 feat: ENG-695 — Embedded Live Browser View in Chat via CDP Screencast
| Contribution | Detail |
|---|---|
CdpClient WebSocket class |
Core client managing the WebSocket connection to Chrome DevTools Protocol |
| Per-task CDP session management | Each task gets its own isolated session — prevents interference between concurrent executions |
| Task lifecycle hooks | CDP sessions open and tear down in sync with task start/stop events |
| Direct CDP session support | Enables direct communication with browser targets without a proxy layer |
| Impact | +895 / −2 lines merged to main — infrastructure powering the live browser view feature |
Maintainer note: "Your PR provided the low-level CDP infrastructure that makes the live browser view possible... Your work (+895 / -2 lines) is now live in main." — @mavishay
| Platform | Handle / Link |
|---|---|
| devparikh-charusat09 | |
| 🐙 GitHub | Dev0907 |
| 📊 Kaggle | molaritydev09 |
| 📖 Springer | Publications |
| 🌐 Lekh.ai | lekhai-frontend.vercel.app |
| 🏓 PicklePro | pickleproproject.vercel.app |
| devparikh200479@gmail.com |
"Code with purpose, build with impact."
⭐ Star my repos if you find them useful!