Naman Siddheshwar

SAT: 1590/1600 (790 Math, 800 English) · Programming: Python, C++, Java, JavaScript · Tools: Node.js, MongoDB, NumPy

• Founded an AI-powered STEM practice platform with Science Bowl simulation, text-to-speech, buzzer practice, and learner analytics, used by students, nonprofits, and local PTAs.
• Designed a solution-centric automatic question generation pipeline for complex STEM and Olympiad problems; submitted research to WAIE.

• Developed CLI Composer, an MCP-based agentic pipeline that generates command-line interfaces from YANG models in a shared production networking codebase.

• Developing a recommender system for older listeners and cognitive engagement using music embeddings, similarity modeling, and DSP-based benchmarks.
• Built a real-time sustain-pedal visualizer using microphone audio, harmonic decay, resonance, and spectral flux to support piano pedagogy.

• Investigated LLM reasoning, designed an experimental evaluation pipeline, and coauthored a paper submitted to a NeurIPS workshop.

• Developed C++ autonomous routines with PID control, motion profiling, and sensor integration; helped the team reach the VEX U.S. Open division final.

• Placed fifth of approximately 80 teams; developed lunar-dust imaging algorithms and evaluated sensor noise and system-design tradeoffs.

• Built an ESP32-controlled robotic piano using solenoids and servos, integrating software, electronics, and mechanics.