NST GenAI Hackathon
Special Jury Award
Wraith OS — RetroCloud
Built a lightweight cloud file-management system with a FastAPI backend.
Hey,It's
OriginalEd. 2026
8.92CGPA
DHRUV
AI / ML Engineer · Researcher
Open
to work
2026
to work
2026
Scroll sidewaysScroll down
Who
dis?
About
I build models that show their work — they cite, abstain, or explain. Never just guess.
I am a Computer Science and Engineering undergraduate at Bennett University (The Times Group), specialising in artificial intelligence, with a CGPA of 8.92. My coursework runs through linear algebra, probability and statistics, machine learning, deep learning, NLP and data structures — and my work sits across all of it.
Most of what I build is about models that can account for themselves. Ruchi verifies every citation in a legal brief against 38,005 Supreme Court judgments. IndicRAG-QA cites a passage or abstains. TruthShield explains its verdicts through attention rollout. On the research side I ran a four-model discriminator ablation to resolve the PatchGAN paradox in MRI tumour segmentation.
B.Tech CSE (AI)
Bennett University
CGPA
8.92
Research Lead
IEEE & ACM
Summer 2026
Publicis Resources
Open to internships✦Agentic AI✦Deep learning research✦RAG that abstains✦PyTorch✦LangGraph✦Rust → WASM✦Bennett University✦Open to internships✦Agentic AI✦Deep learning research✦RAG that abstains✦PyTorch✦LangGraph✦Rust → WASM✦Bennett University✦
Selected work · 6 packs
The
goods
Research ablations, agents that ground every answer, and systems code compiled to run in the browser.
Full index →The workshop · interactive
Step into
the room
Welcome to the workshop
Click anything in the room.
Every object opens part of the story — or pick from the list.
Receipts,
not
adjectives
38,005
Supreme Court judgments Ruchi checks citations against
0.958
SSIM — best PatchGAN in the BraTS ablation
54.1%
Recall on unseen attacks, up from 8.5%
0.9956
Validation F1 — TruthShield's fusion head
132
Tests across the Rust systems projects
8.92
CGPA — B.Tech CSE (AI)
The
record
Jun 2026 — Aug 2026
Experience
Applications Admin Intern — AI/ML
Publicis Resources
Gurugram, India
- Administered and tested AI-enabled enterprise applications in production.
- Handled DSAR workflows and data validation; documented processes for MLOps teams.
2025 — 2026
Experience
Research Lead
IEEE & ACM Student Chapters, Bennett University
- Leading student AI/ML research across both chapters.
2024 — 2028
Education
B.Tech — Computer Science & Engineering (Specialisation in AI)
Bennett University, The Times Group
Greater Noida, India
- CGPA 8.92.
- Relevant coursework: Linear Algebra, Probability & Statistics, Machine Learning, Deep Learning, NLP, DSA.
Technical skills
Toolkit
ML / DL
- PyTorch
- TensorFlow
- Hugging Face Transformers
- Scikit-learn
- OpenCV
- NumPy
- Pandas
Models
- U-Net
- Conditional GANs (PatchGAN)
- BERT / RoBERTa
- GPT-2
- Vision Transformers
- Autoencoders
GenAI & Agents
- LangGraph
- LangChain
- Strands Agents SDK
- LiteLLM
- AWS Bedrock
- RAG
- Tool Calling
Languages
- Python
- Rust
- TypeScript / JavaScript
- SQL
Systems & Deployment
- FastAPI
- Docker
- Modal
- Vercel
- WebAssembly
- Tokio
- Rayon
- SQLite
- Redis
- Git
Some
silver
ware
Project Showcase 2025 — Bennett University × Times of India
2nd Runner-Up
TerraVit
Built an AI climate-intelligence system that applies Vision Transformers to satellite imagery.
Fresh off GitHubpenumbra-nidsPython↗Fresh off GitHubprofileTypeScript↗Fresh off GitHubIndicRAG-QAPython↗Fresh off GitHubRuchiPython↗Fresh off GitHubtruthshield_vfJupyter Notebook↗Fresh off GitHubRust-projects-deployment-Rust↗
Let's
work!
Open to internships & research roles
Internships, research collaborations, or an agent that needs to stop making things up — write to me.
guptadhruv1402@gmail.comBack to start