const mayank = {
  location   : "IIT Jodhpur, Rajasthan 🇮🇳",
  degree     : "B.Tech — Computer Science & Engineering (2023–27)",
  cgpa       : 7.52,

  focus      : ["Backend Development", "AI/ML Systems", "Full-Stack Engineering"],
  currently  : ["Building scalable web platforms", "Exploring deep learning architectures",
                "Sharpening DSA with C++"],

  superpowers: {
    algorithms : "C++ · 400+ DSA problems · 150+ LeetCode",
    backend    : "Node.js · REST APIs · MongoDB",
    ai_ml      : "CNN · PCA · Scikit-learn · OpenCV",
    research   : "NLP pipelines · Audio synthesis · Image clustering",
  },

  jee        : { advanced: "AIR 2035 (Top 0.2%)", main: "AIR 7302 (Top 0.6%)" },
  contact    : "mayankdahiya8431@gmail.com",
  motto      : "Ship things. Learn fast. Build what matters.",
};

Undergraduate Researcher — IIT Jodhpur  Jan 2025 – Apr 2025

Under Prof. Romi Banerjee, CSE Dept.

• Engineered a multi-stage pipeline to analyze comic panels, extract text, and synthesize synchronized audio narrations — achieving 95% accuracy.
• Programmatically merged 50+ audio clips into a coherent, seamless final narrative.

Full-stack platform for managing coding and non-coding assignments with automated evaluation.

• Handles 100+ concurrent users with real-time, sandboxed code execution via JDoodle API.
• Instructors can upload 50+ problems with test cases; students get instant graded feedback.
• Stack: Node.js · React.js · MongoDB · Tailwind CSS

Custom KD-Tree implementation for high-dimensional image similarity search in C++.

• Built a KD-Tree from scratch for fast indexing of 300 high-dimensional vectors.
• Optimized nearest-neighbor search reducing query time by up to 84% vs. linear scan.
• Stack: C++

End-to-end image classification pipeline on CIFAR-10.

• Applied PCA (90% variance retained) and benchmarked classical ML baselines.
• Designed and trained a custom CNN from scratch, achieving 80% accuracy on CIFAR-10.
• Stack: Python · NumPy · Scikit-learn · Matplotlib

Numerical solver built entirely in C — no external math libraries.

• Built a matrix parser and recursive determinant calculator supporting up to 10×10 systems.
• Validated against a full test suite — 100% computational accuracy.
• Stack: C

 

• Assistant Head, E-Cell — Directed entrepreneurship workshops for 200+ students; drove increased participation in campus startup initiatives.
• Team Coordinator, Esports Club — Managed BGMI tournaments for 50+ participants; led regular practice sessions.
• Assistant Head, Ignus — Led a Krafton-sponsored gaming event; coordinated logistics, promotions & match operations for 300+ participants.

"Engineering solutions that speak for themselves."

Open to internships, research collabs, and dev opportunities — let's build something impactful!