Raj Patel — AI Engineer & Full-Stack Developer in Ahmedabad, India

Raj Patel

AI Engineer & full-stack developer — alumnus of the Intel® AI for Manufacturing and Microsoft Elevate × GTU internships, currently available for work.

Focused on agentic AI, machine learning & data science. Building from Ahmedabad, India.

About me

I build agents, train models, ship products.

Hey! I'm Raj — I love turning ideas into real AI-driven products and finding smarter, faster ways to improve systems. My work lives where machine learning meets working software: agentic pipelines, computer-vision models and web apps people actually use.

I graduated in Information Technology from SAL Engineering & Technical Institute (Gujarat Technological University) with an 8.77 CGPA, alongside a Minor/Honours in Data Science — GTU's four-semester advanced track, one extra subject every semester, and I qualified for all of it. My final year closed with a 12-week Microsoft Elevate × GTU internship (Jan–Apr 2026), where NeuroAssist was my capstone.

And I vibe-code full products end to end: Django backends, Streamlit tools, REST APIs, dashboards — shipped and deployed, not just committed. Always curious, always learning, always levelling up.

Agentic AI 01

Agentic systems, RAG pipelines, LLMs & SLMs, prompt & tool-use orchestration, HuggingFace, Claude Code.

AI Engineering 02

Python, PyTorch, TensorFlow, scikit-learn, NLP, YOLOv8, OpenCV, 3D Attention U-Net, feature engineering, model deployment.

Full-Stack · Vibe Coder 03

Django, Node.js, JavaScript, HTML/CSS, AngularJS, Streamlit, REST APIs, MySQL, MongoDB, PostgreSQL.

Data Science & Cloud 04

Pandas, NumPy, Matplotlib, BigQuery, Google Cloud Platform, Jupyter, Kaggle, Google Colab, data storytelling.

The toolbox

My Skills

Scroll — the toolbelt glides past

PythonPythonLanguage
JavaScriptJavaScriptLanguage
CCLanguage
HTML5HTML5Language
CSSCSSLanguage
PyTorchPyTorchAI & ML
TensorFlowTensorFlowAI & ML
scikit-learnscikit-learnAI & ML
OpenCVOpenCVAI & ML
NumPyNumPyAI & ML
pandasPandasAI & ML
Agentic AIAI & ML
RAGRAGAI & ML
Y8YOLOv8AI & ML
MatplotlibAI & ML
GPTChatGPTAI Tools
ClaudeClaudeAI Tools
Google GeminiGeminiAI Tools
DjangoDjangoWeb & Frameworks
FlaskFlaskWeb & Frameworks
StreamlitStreamlitWeb & Frameworks
Node.jsNode.jsWeb & Frameworks
AngularAngularJSWeb & Frameworks
MySQLMySQLDatabases
MongoDBMongoDBDatabases
PostgreSQLPostgreSQLDatabases
Google BigQueryBigQueryDatabases
Google CloudGoogle CloudCloud & Data
Google ColabGoogle ColabCloud & Data
KaggleKaggleCloud & Data
Hugging FaceHugging FaceCloud & Data
JupyterJupyterIDE
AntigravityIDE
VS CodeIDE
{ }CodexIDE
Claude CodeClaude CodeIDE
GitGitVersion Control
GitHubGitHubVersion Control

Proof, not promises

The Vault

1 / 7 — keep scrolling

Featured work

My Projects

From production CV systems to agentic RAG pipelines — every build here solves a real-world problem, end to end.

NeuroAssist

My final-year capstone under the Microsoft Elevate Program: brain-tumor detection & clinical decision support for high-grade gliomas — a 3D Attention U-Net trained on BraTS 2020 across four MRI modalities, plus RAG-enhanced LLM reasoning drafting structured, WHO-aligned diagnoses.

TensorFlow3D Attention U-NetRAG + LLMBraTS 2020
neuroassist — inference
$ python neuroassist/infer.py --scan patient_042.nii.gz
loading 3D Attention U-Netok
multi-modal MRI: T1 / T1ce / T2 / FLAIR (BraTS 2020)
segmenting tumor sub-regions … done (41.2s)
dice — whole 0.91 · core 0.87 · enhancing 0.84
RAG: retrieving WHO CNS5 references … 6 passages
LLM report: WHO-aligned draft → report_042.pdf
clinical summary ready for review

BraTS2020 · MGMT

A deep-learning pipeline for brain-tumor analysis on multi-modal MRI: preprocessing T1, T1CE, T2 & FLAIR into NumPy arrays, an interactive slice-visualiser to validate it, and an MGMT promoter-methylation predictor on a custom 3D ResNet. Preprocessed dataset published on Kaggle.

PyTorch3D ResNetBraTS2020Kaggle
Open the Kaggle dataset ↗
brats2020 — mgmt pipeline
$ python preprocess.py --brats2020 --modalities T1,T1CE,T2,FLAIR
resampling volumes → .npy arrays … ok
slice visualiser: validating anatomy … passed
$ python train_mgmt.py --arch resnet3d --epochs 60
task: MGMT promoter methylation (binary)
train loss falling · val AUC improving …
dataset published → kaggle.com/raj072004/preprocessed-final

PCB Inspection AI

Industry-grade optical PCB defect detection for smartphone assembly, built during the Intel internship. YOLOv8 at mAP@0.5 ≈ 0.90 across 10 defect classes, with real-time webcam & batch inspection via Streamlit.

YOLOv8OpenCVStreamlitIntel AI
Read the internship report ↗
pcb-inspection — realtime
$ streamlit run app.py --source webcam
loading YOLOv8 weights … ok
classes: 10 defect types (short, spur, open, …)
stream: 1080p @ 24fps — live overlay on
frame 0412 — spurious_copper 0.93 · mouse_bite 0.88
frame 0413 — clean
mAP@0.5 ≈ 0.90 across all classes
batch report saved → inspection_run_07.csv

Hospital MS

A Django web app streamlining hospital operations — patients, appointments, billing, pharmacy — with role-based staff access, PDF reports and secure auth. 88★ on GitHub and counting.

DjangoMySQLRBAC88★ GitHub
hospital-management-system-yhsp.onrender.com
Hospital Management System — live Django app

SAC AI Search

Open-source GST SAC-code database with AI search: describe any service in plain language and it finds the right Service Accounting Code — NLP matching in place of manual lookup through thousands of codes.

PythonNLPSemantic SearchOpen Source
sac-ai-search
$ sacfind "monthly bookkeeping and GST filing for a bakery"
embedding query … ok
searching SAC database …
1. 998222 — Accounting & bookkeeping 0.94
2. 998231 — Tax consultancy (GST) 0.91
3. 998399 — Other professional svcs 0.62
best match: SAC 998222

Reconciliation

Wallet-balance reconciliation system with an atomic compute-then-commit engine — Flask dashboard, CLI and batch runner backed by MS SQL Server, built for zero-drift financial integrity.

FlaskMS SQL ServerFintechCLI
reconciliation — batch runner
$ recon run --batch 2026-07 --engine atomic
wallet ledger ← MS SQL Server … 48,112 rows
compute phase: balances staged & hashed
commit phase: all-or-nothing
matched 47,893 · flagged 219 · drift ₹0.00
batch reconciled → dashboard updated

ExpenseFlow

An AI expense tracker that combines receipt capture, smart extraction, spend insights and searchable history in a polished responsive workspace — from receipt to saved expense in under 3 minutes.

JavaScriptAI ExtractionVercel
expenseflow-roan.vercel.app
ExpenseFlow — AI expense tracker

Open to work. Let's build.

I'm looking for AI Engineering, Agentic AI and Data Science roles — and I take on freelance builds. From fine-tuning models to shipping the whole product, I go from idea to deployed, fast.

Contact me

Milestones

Proof of work — on paper and in production

Internship

Intel® AI for Manufacturing — Intel × GTU × GCCI

Part of the first batch of Intel Corporation's first-ever university summer internship & upskilling program with GTU and GCCI — 16 weeks (Feb – Jul 2025) building production-grade AI for manufacturing.

View certificate ↗
Recognition

Invited to present to GTU's Vice Chancellor & Intel leadership

Showcased our project to the VC of GTU, Intel's Global Director for Digital Readiness and Intel's Senior Director for Asia Pacific & Japan.

See the LinkedIn post ↗
Internship

Microsoft Elevate × GTU — 12-week AI/ML internship

Final-year internship (Jan – Apr 2026) under the Microsoft Elevate Program, facilitated by FICE Education — built NeuroAssist, my AI cancer-detection capstone, end to end.

View certificate ↗
Education

B.E. in Information Technology — 8.77 CGPA

SAL Engineering & Technical Institute, Gujarat Technological University, Class of 2026.

Honours

Minor/Honours in Data Science — GTU advanced track

One extra subject per semester for four semesters, on top of the core IT degree — qualified for the full track.

Open Source

88★ on GitHub

The Hospital Management System is my most-starred repo — real users, real issues, real maintenance.

Language

IELTS 7.0 — CEFR C1 English

Overall band 7.0, mapped to CEFR C1 — comfortable working, writing and presenting in English for global teams.

View score report ↗

For Any Questions

The questions recruiters and collaborators ask most

Profile

Who is Raj Patel?

Raj Patel is an AI Engineer and Full-Stack Developer in Ahmedabad, Gujarat, India. He builds across agentic AI, machine learning and deep learning with PyTorch and TensorFlow, NLP and LLMs, computer vision and data science, and is completing a B.E. in Information Technology at Gujarat Technological University in the Class of 2026.

Availability

Is Raj Patel available for hire?

Yes. Raj is available as an AI Engineer and Full-Stack Developer in Ahmedabad, and works remotely with teams elsewhere in India and abroad. He is open to AI Engineering, Agentic AI, Machine Learning and Data Science roles, and also takes on freelance builds. The fastest way to reach him is email at rajpatel9408019@gmail.com or via LinkedIn.

AI / ML

What AI, machine learning and deep learning work does Raj Patel do?

Raj builds deep-learning models in PyTorch and TensorFlow — 3D segmentation networks for medical imaging, YOLOv8 detection for manufacturing defect inspection — alongside agentic AI and retrieval-augmented generation pipelines over LLMs. On the data science side he covers preprocessing, feature engineering, evaluation and deployment, so an AIML project ships as a working product rather than a notebook.

Work

What is Raj Patel's strongest project?

NeuroAssist, his final-year capstone under the Microsoft Elevate Program. It performs brain-tumor detection and clinical decision support for high-grade gliomas using a 3D Attention U-Net trained on BraTS 2020 across four MRI modalities (T1, T1ce, T2, FLAIR), with a RAG and LLM layer that drafts WHO CNS5-aligned clinical reports.

Experience

What internships has Raj Patel completed?

Two, both through GTU. Intel's AI for Manufacturing with GTU and GCCI was his summer internship after semester 6 — it counted as GTU's 15-day summer internship, but he ran it for a full 16 weeks (February to July 2025) as part of the program's first batch. The Microsoft Elevate Program with GTU was his final 8th-semester internship, a 12-week AI/ML placement from January to April 2026.

Stack

What technologies does Raj Patel work with?

Python, PyTorch, TensorFlow, scikit-learn, OpenCV, YOLOv8, HuggingFace and agentic RAG pipelines on the AI side; Django, Flask, Streamlit, Node.js and AngularJS on the web side; and MySQL, MongoDB, PostgreSQL, BigQuery and Google Cloud for data and infrastructure.

Education

What has Raj Patel studied?

A B.E. in Information Technology at SAL Engineering & Technical Institute, Gujarat Technological University, graduating in the Class of 2026 with an 8.77 CGPA. He also qualified for GTU's advanced track, taking a Minor/Honours in Data Science with one extra subject per semester for four semesters.

Contact

Say hello.

Raj Patel AI Engineer & Full-Stack Developer

Got a role, a project, or a wild AI idea? My inbox is open — I usually reply within a day.

Based in
Ahmedabad, Gujarat, India
Instagram

Opens in your email app, pre-filled — nothing is stored anywhere.