👋 Hi, I’m Samiksha Mitra — an undergraduate at IIT Guwahati deeply driven by Quantitative Finance, Machine Learning, and Applied AI.
I’m passionate about transforming mathematical ideas into high-performance systems that operate under real-world constraints. My work spans ML-based trading strategy research, signal modeling, backtesting, and risk evaluation, along with applied projects in fraud detection, computer vision for robotics, and scalable software systems.
I enjoy diving deep into data, models, and systems, constantly experimenting, validating assumptions, and refining ideas through rigorous evaluation. I’m especially motivated by problems at the intersection of finance, AI, and software engineering, where precision, robustness, and impact truly matter.
🤖 Fraud Detection System
GitHub
- Detected high-risk fraudulent transactions while minimizing false positives in imbalanced financial data.
- Built and tuned an ML-based detection pipeline using EDA and business-aware evaluation metrics.
- Gained insight into cost-sensitive modeling, precision–recall trade-offs, and explainable ML systems.
🌊 Antahsagari – Underwater ROV for Control & Navigation
GitHub | Website
- Enabled reliable real-time vision for underwater navigation and object detection.
- Designed and deployed optimized vision pipelines on embedded hardware under latency and visibility constraints.
- Learned system-level trade-offs between model complexity, speed, and robustness in deployed robotics systems.
🪐 Martian Core Analysis
GitHub
- Analyzed subsurface anomalies in Martian core structures using seismic data.
- Generated synthetic datasets and combined ML models with physics-informed constraints.
- Learned how incorporating domain knowledge improves ML robustness and interpretability.
GitHub | Website
- Built a scalable and user-friendly airline management platform.
- Designed an end-to-end web system focusing on clean UX, dynamic workflows, and backend scalability.
- Learned how frontend UX decisions and backend architecture directly impact product usability and adoption.