class AmanSagar:
def __init__(self):
self.name = "Aman Sagar"
self.degree = "Bachelor's in Data Science & Applications — IIT Madras (GPA: 8.01)"
self.currently = "Intern @ Vicharanashala Lab · IIT Ropar"
self.focus = ["ML Research", "LLM Systems", "Data Analysis"]
self.open_to = ["Research Roles", "ML Internships", "AI Collaborations", "Data Analyst"]
def highlights(self):
return [
"100% recall on breast cancer classifier — zero false negatives",
"Top 15 · Annam.ai National Soil Classification Challenge",
"Top 10 Finalist (Rank 8) — Design-a-Thon 2025, IIIT Nagpur",
"Google GenAI Exchange Hackathon — Selected Finalist",
"Built LLM Agents, RAG systems, and clinical AI tools from scratch",
]I'm a Data Science Student and ML Practitioner who builds things that work in the real world. My work spans clinical ML, deep learning for agriculture, LLM-powered applications, and production-grade full-stack systems. I care deeply about interpretability, deployment, and real impact.
Languages
ML / Deep Learning
AI / LLM Systems
Web Development
Databases & Tools
|
Research · IISER Bhopal Clinically-driven ML pipeline built to minimize false negatives in cancer detection. Applied cost-sensitive learning (Fβ ≈ 7.8), achieving 100% malignant recall with zero false negatives. Includes SHAP interpretability, reject-option classifier, and a deployed Streamlit decision support app.
|
Annam.ai · IIT Ropar · Top 15 🏆 Deep learning models for rice & wheat disease classification using EfficientNetV2B3 and ResNet50. Built large-scale annotated datasets with CLAHE preprocessing pipelines and Test Time Augmentation. Ranked Top 15 in national soil classification challenge.
|
|
Personal Suite · 2025 A collection of production-focused LLM tools: an autonomous Multi-Step Agent with tool usage, a Virtual Teaching Assistant powered by RAG, a Data Analyst Agent (natural language → insights), and an automated PPT Generator.
|
Zinnovatio 3.0 Hackathon · Top 50 Team 🚀 Full-stack AgriTech platform with market price dashboard, AI-based crop advisory, disease detection, and a farmer community forum. LLM-powered advisory flows. Also shortlisted at SuperHack 2025.
|
|
Full-Stack · Production-Grade Complete hospital backend with appointment scheduling, user authentication, role-based access, and async background job processing using Celery and Redis. Clean Vue.js frontend.
|
Power BI Analytics Interactive Power BI dashboards analyzing EdTech course trends (learner engagement, instructor impact, subtitle effects) and Netflix content catalog (genre trends, regional production, release patterns).
|
| Achievement | Event | Year | |
|---|---|---|---|
| 🥇 | Top 10 Finalist · Rank 8 | Design-a-Thon 2025 · IIIT Nagpur (SoulSync — AI wellness assistant) | 2025 |
| 🤖 | Selected Finalist | Google GenAI Exchange Hackathon (AI Marketplace Assistant) | 2025 |
| 🌱 | Top 15 | Annam.ai Soil Classification Challenge Parts 1 & 2 | 2025 |
| 🚀 | Top 50 Qualified Team | Zinnovatio 3.0 Hackathon (KrishiSetu) | 2025 |
| ⚡ | Shortlisted | SuperHack 2025 (KrishiSetu Prototype) | 2025 |
| 🎖️ | Participant | Smart India Hackathon (VighnaSetu) | 2025 |
- 🔬 Building interpretable ML pipelines for clinical applications
- 🤖 Exploring advanced RAG architectures and LLM agent frameworks
- 📊 Deepening expertise in MLOps and model deployment at scale
- 🎓 B.S. Data Science & Applications — IIT Madras (Expected 2027)
- 📝 Interning at Vicharanashala Lab · IIT Ropar