🎙️
Click the button and speak
Try: "What projects has Aditi built?" or "Is she open to jobs?"
aditi@portfolio ~ $
Developer Intern @ Soft Solvers · open to jobs & internships
AI Engineer & Full Stack Developer

Aditi
Raj

B.Tech CSE — AI & ML  ·  Vellore Institute of Technology  ·  2027

I build AI systems that ship to the real world — not notebooks, not demos. From breast cancer detection running offline on Android, to a generative fashion system on Stable Diffusion, everything goes end-to-end: model training, backend APIs, and production deployment. Right now I'm a Developer Intern at Soft Solvers, automating insurance test flows with Playwright; before that I worked on AI agents at TroviqAI.

📍 Dhanbad, Jharkhand
Aditi Raj speaking at VIT
Speaking at Vellore Institute of Technology
Aditi Raj
8.3/10
CGPA · VIT
400+
DSA problems solved
92%
Peak model accuracy
2
S-Grade projects
Python·TensorFlow·React.js·Stable Diffusion·FastAPI·React Native·PyTorch·CLIP·MongoDB·GCP·Docker·Scikit-Learn·Node.js·Firebase·Android Studio·XGBoost·EfficientNetB0·PostgreSQL·Computer Vision·Generative AI·LangGraph·Playwright·DINOv2·RAG·Grad-CAM·EasyOCR    Python·TensorFlow·React.js·Stable Diffusion·FastAPI·React Native·PyTorch·CLIP·MongoDB·GCP·Docker·Scikit-Learn·Node.js·Firebase·Android Studio·XGBoost·EfficientNetB0·PostgreSQL·Computer Vision·Generative AI·LangGraph·Playwright·DINOv2·RAG·Grad-CAM·EasyOCR    
📍 Dhanbad, Jharkhand $ ./easter_egg.sh
About

Built for the real world,
not the notebook.

📍 From Dhanbad, Jharkhand
🎓 Vellore Institute of Technology
💼 Developer Intern, Soft Solvers Solutions
🤖 Main Council Member, AI Club — VIT
📅 B.Tech CSE (AI & ML), Class of 2027
🏅 400+ DSA problems · 2 S-Grade projects
🥇 Rank 1, IEEE Summer of Code 2026
Aditi Raj at VIT, wearing her AI Club ID
At VIT with my AI Club ID

I grew up in Dhanbad, Jharkhand, and now study at Vellore Institute of Technology — pursuing a B.Tech in Computer Science with a specialisation in AI and Machine Learning.

I started writing code because I wanted to solve things that mattered — diagnosing crop disease without internet, detecting breast cancer early on a cheap Android phone, catching e-commerce fraud at scale. My technical range spans the full ML stack: dataset curation, model training, TFLite + XGBoost inference, FastAPI backends, React Native frontends, and production deployment.

I was promoted from Core Member to Main Council Member of the AI Club at VIT, where I've led workshops on LLMs and deep learning for 300+ students and helped organise hackathons and speaker sessions. I was also selected for VIT's Experiential Learning Program, visiting BSNL, T-Hub, T-Works, Zensar Technologies, and L&T Metro in Hyderabad and Pune.

Lately my work has moved toward AI systems you can trust: models that say when they're unsure (SkinIQ), agents behind guardrails and human approval (AgentHub, and the Hunt agent at TroviqAI), and test suites that run themselves (328 Playwright tests at Soft Solvers). I was also selected for Amazon ML Summer School 2026, one of about 3,000 from 150,000+ applicants.

Projects

Shipped work — not just ideas

SkinIQ landing page6 screens
🧴
New · Multi-agent
SkinIQ
Skincare guidance that shows its work. Four agents (vision, ingredient check, recommendations, symptom check) read a face photo and product labels. A DINOv2 model gives tone, type, acne and dark-circle estimates with real confidence ranges, and falls back to your own input when unsure. EasyOCR checks every ingredient against a 28,000-name database.
4 agents · 28K ingredients · conformal prediction
PyTorchDINOv2EasyOCRFastAPIGroqPostgreSQL
AgentHub blocking a prompt-injection attempt3 screens
🛂
New · Agent security
AgentHub
Five AI agents behind one hardened pipeline: API-key gateway with rate limiting, a three-layer prompt-injection check (regex, a local DeBERTa classifier, NeMo Guardrails locally), LangGraph routing, PII redaction, and password-gated human approval for any tool that changes data. Every request is traced stage by stage.
5 agents · 3-layer injection defence · RAGAS-scored RAG
FastAPILangGraphGroq · LiteLLMHF TransformersRAGASPostgresReact + Vite
▶ Demo ↗
browser ─▶ forward proxy (Python) ─▶ internet
  reads each request's domain (HTTP / HTTPS · TLS · TCP)
allowlist + age tier ─▶ allowed ✓ / blocked ✗
voice question ─▶ TF-IDF over your notes ─▶ guiding reply
📚
Live · Deployed
StudyFocus
A study platform with its own Python forward proxy: it handles HTTP and HTTPS traffic over TCP sockets, reads each request's destination domain, and only lets allowlisted sites through during study time, with stricter limits for younger students. A voice tutor answers in English, Hindi or Hinglish from the student's own notes and asks guiding questions instead of giving answers away.
Computer networks · TF-IDF RAG · EN / HI / Hinglish
PythonFastAPIHTTP/HTTPSTLSTCP socketsTF-IDFWebSockets
ShieldMart Grad-CAM damage heatmap4 screens
🛡️
Live · Deployed
ShieldMart / Returns Shield
India's first AI-verified used goods marketplace. XGBoost fraud detection (ROC-AUC 0.91, 590K transactions) + EfficientNetB0 damage detection with Grad-CAM heatmaps. Built for Flipkart Grid 6.0.
ROC-AUC 0.91 · 590,540 training samples · 4 damage types
React 18FastAPIXGBoostEfficientNetB0PostgreSQLRazorpayDocker
🌾
Featured
CropDoc + StreetCare
AI mobile app for crop disease diagnosis and street animal healthcare. Offline-first with SMS fallback and multilingual voice — designed for rural India with no internet dependency.
87.4% accuracy · 38 disease classes · PlantVillage
React NativeTFLiteMobileNetV3FastAPISupabase
👗
Generative AI
MoodBoard to Outfit
End-to-end GenAI fashion system — converts mood descriptions to photorealistic outfit images via CLIP embeddings + Stable Diffusion v1.5 + IP-Adapter FaceID virtual try-on.
25–40 sec generation · Public Colab URL
CLIPStable Diffusion v1.5IP-AdapterPyTorchGradio
🩺
S Grade · Highest Distinction
BreastScan
Android app for early breast cancer detection. On-device ML inference — sensitive data never leaves the phone. OCR extracts clinical values from medical reports automatically.
90%+ accuracy · Random Forest · 5-fold CV
Android StudioJavaTFLiteML Kit OCRFirebase
🥗
S Grade · Highest Distinction
NutriScan
Food grading system classifying packaged products A–E by nutritional value. Custom 750+ product dataset built from scratch. Full model comparison with precision, recall, F1 visualisations.
92% accuracy · 750+ products · Random Forest
PythonScikit-LearnPandasRandom Forest
🍕
Live · Production
Pizza Delivery App
Production MERN platform with Razorpay payments (HMAC-SHA256), JWT auth, live order tracking, admin dashboard, CI/CD via Vercel + Render + MongoDB Atlas.
Deployed · Vercel + Render + MongoDB Atlas
React.jsNode.jsMongoDBRazorpayJWT
How it works

Screen by screen, step by step

Skills

Technical toolkit

AI & Machine Learning
TensorFlow / TFLite90%
Scikit-Learn / XGBoost88%
Computer Vision / CNNs85%
PyTorch82%
Generative AI (SD, CLIP)78%
Full Stack
React.js / React Native87%
Node.js / Express.js83%
FastAPI / Python Backend85%
MongoDB / Firebase / Supabase80%
All Technologies
PythonC++JavaJavaScriptTypeScriptSQLTensorFlowPyTorchXGBoostScikit-LearnCLIPStable DiffusionEfficientNetB0Grad-CAMReact.jsReact NativeNext.jsNode.jsFastAPIExpress.jsMongoDBPostgreSQLFirebaseSupabaseGCPDockerVercelRenderGitHub ActionsAndroid StudioGradioRazorpayLangGraphLangChainRAGFAISSTF-IDFDINOv2EasyOCRGroqClaude APIAmazon BedrockLLM GuardrailsPlaywrightPandasNumPyAWSWebSockets
Certifications
🔵
Google IT Support Professional
Google
🟠
Cloud Computing
NPTEL · IIT Kharagpur · Elite 70%
🟢
Google Cloud Skill Badges
Google Cloud
🟣
Internet of Things
NPTEL · IIT Kharagpur · Elite 87%
🔷
Bits & Bytes of Networking
Google
🎓
Web Dev Bootcamp
Udemy · 2024
〽️
Applied Machine Learning in Python
Univ. of Michigan · Coursera
☁️
AWS Technical Essentials
AWS Training & Certification
👁️
Computer Vision
VITyarthi · VIT
📊
GenAI Data Analytics Job Simulation
Tata · Forage
🌐
Networking Basics
Cisco Networking Academy
🗄️
Advanced SQL for Interviews
Bosscoder Academy
Experience

Where I've been

Sep 2026 – Now
Gurugram
Developer Intern Current
Soft Solvers Solutions Pvt. Ltd. · Gurugram, India
  • Working on the Playwright test automation project for Agiliux Insurance, turning test scenarios from business analysts into automated scripts.
  • Wrote scripts that generate a structured, client-facing test report every time the full pipeline runs (modules, test cases, results, open issues). The suite currently has 328 passing test cases.
  • Studied the existing test cases for the golden instance and the INS instances, and wrote new ones on top of them.
  • The automation fills in and checks long insurance forms that testers used to fill in by hand.
Offer email: selected as Developer Intern at Soft Solvers Solutions
Offer: Developer Intern, Soft Solvers (Sep 2026)
Project roles table: Aditi assigned to the Playwright (RI) project
Assigned to the Playwright (RI) project
Second CRM employee dashboard: Aditi, Developer Intern
My Second CRM employee profile
PlaywrightTest automationTest reporting
Sep 2026
Remote
Agentic AI Intern
Yogam Venture Studio · TroviqAI · troviqai.com ↗
  • Contributed to the TroviqAI website and worked on the Hunt agent, one of the platform's B2B marketing agents built as a LangGraph workflow.
  • Built the Solo Agent Coach feature: a coach writes one sentence about how they work, adds a website or resume PDF (or answers a few questions), and the platform drafts their portfolio story. Wrote the init SQL scripts that create its PostgreSQL tables.
  • Worked on and debugged the Hunt agent's safeguards: token-bucket rate limiting per organisation and per vendor API, per-call token limits, PII redaction, and blocking off-topic or prompt-injection inputs.
Welcome onboard email from Yogam Venture, payment amounts hidden
Welcome email from Yogam Venture
Solo Agent Coach personalize step
Solo Agent Coach · personalize step
Portfolio drafted by Solo Agent Coach
The portfolio it drafts
LangGraphFastAPIPostgreSQLReactRate limitingLLM guardrails
Jul – Aug 2026
Remote
Amazon ML Summer School 2026
Amazon · selected as one of ~3,000 from 150,000+ applicants
Cleared a technical screening in DSA, linear algebra and probability, then studied core machine learning algorithms from basic to intermediate level and used them in my projects.
May – Aug 2026
Software Engineering & AI Foundation Intern
MPOnline Limited · JV of Govt. of Madhya Pradesh & TCS
Built a MERN food-delivery app with JWT login, Razorpay payments and Google Maps.
2025 – Now
Main Council Member — AI Club
Vellore Institute of Technology · promoted from Core Member
Led technical workshops on LLMs and deep learning for 300+ students, and helped organise campus hackathons, technical events and speaker sessions.
Oct–Nov 2025
1 month
Experiential Learning Program — Industrial Training
VIT · Hyderabad & Pune
Selected for VIT's flagship program. Visited BSNL, T-Hub, T-Works, Zensar Technologies, and L&T Metro, seeing how telecom, hardware prototyping, IT services and startup incubation teams work.
Aditi Raj at T-Works, Hyderabad
T-Works, Hyderabad
Aditi Raj during the industrial visit in Hyderabad
Industrial visit, Hyderabad
2023 – 2027
B.Tech — Computer Science (AI & Machine Learning)
Vellore Institute of Technology · CGPA 8.3 / 10
Machine Learning · Deep Learning · NLP · Computer Vision · Reinforcement Learning · DSA · DBMS · OS · Computer Networks · Cloud Computing
Achievements
Rank 1
IEEE Summer of Code 2026, a national open-source fellowship
Zonal Finalist
Women Who Master Hackathon 2026, built a GenAI virtual try-on app
Amazon MLSS
One of ~3,000 selected from 150,000+ applicants for ML Summer School 2026
400+
DSA problems solved on LeetCode and GFG
2 × S Grade
Highest grade (90%+) in two VIT project courses
Flipkart GRiD
Cleared GRiD rounds; ShieldMart was built for GRiD 6.0
AI Bot

Ask AditiBot — or match your JD

Ask the bot anything about my work, or paste a job description to see which of its skills I already have. AditiBot runs on Groq, and if it can't be reached it still answers from my profile.

Mode: Groq

AditiBot
Groq
Her projects ShieldMart ML skills Soft Solvers TroviqAI SkinIQ AgentHub Availability
Hi! I'm AditiBot. Ask me anything about Aditi's internships, projects or skills.
JD Match Analyzer
Instant Score

Paste any CS/tech job description. You'll see the technologies it asks for that Aditi has used, and the ones she hasn't yet.

—
Match score
vs Aditi's profile
Contact
Aditi Raj
Aditi RajDeveloper Intern @ Soft Solvers
B.Tech CSE (AI & ML) · VIT 2027

Let's build
something real.

I'm open to full-time jobs and internships in software development, full stack, AI/ML and AI agents, as well as research collaborations, and open-source AI projects — especially in computer vision, LLM agents, generative AI, or full stack.

✓ Opening your mail client…