Projects
Creative Works
Projects
A collection of applications and projects I've built, exploring the intersection of design, technology, and user experience.
GPU Memory Calculator
AI/ML · Web Tool
Calculate GPU memory requirements for training Large Language Models with support for PyTorch DDP, DeepSpeed ZeRO, Megatron-LM, and FSDP.
AI Interview Prep
AI/ML · Study Tool
A local AI interview preparation tool for AI research scientist and engineer roles. Practice, learn, and track your progress across key ML/AI domains.
My iOS App
iOS · Featured
A beautifully crafted iOS application built with SwiftUI, featuring intuitive design and smooth user experience.
GPU Memory Calculator
AI/ML · Web Tool
A comprehensive Python application for calculating GPU memory requirements for Large Language Model (LLM) training and inference. The calculator supports multiple distributed training engines including PyTorch DDP, DeepSpeed ZeRO (stages 1-3), Megatron-LM, Megatron+DeepSpeed, and PyTorch FSDP. It also supports inference engines like vLLM, TGI, TensorRT-LLM, and SGLang with KV cache optimization.
Key Features
Training Memory Calculation
- • Multiple Training Engines: PyTorch DDP, DeepSpeed ZeRO, Megatron-LM, and FSDP
- • Detailed Breakdown: Memory by component (parameters, gradients, optimizer states, activations)
- • Preset Models: Quick-load configurations for LLaMA 2, GPT-3, Mixtral, and more
- • Advanced Features: CPU/NVMe offloading, activation checkpointing, multi-node training
Inference Support
- • Multiple Engines: HuggingFace Transformers, vLLM, TGI, TensorRT-LLM, SGLang
- • KV Cache Optimization: Quantization options (INT8, FP8, INT4)
- • Tensor Parallelism: Automatic memory distribution across GPUs
- • Performance Metrics: Throughput estimation (tokens/second)
Interfaces & Tools
- • Web Interface: Interactive UI with real-time validation and formula explanations
- • Command Line: CLI tool for quick calculations and batch processing
- • Python API: Programmatic access for integration into workflows
- • Framework Export: Generate configs for Accelerate, Lightning, and Axolotl
AI Interview Prep
AI/ML · Study Tool
Preparing for AI research scientist and machine learning engineer interviews is uniquely challenging — the field evolves rapidly, questions span deep theory and practical systems, and there's no single resource that ties it all together. This tool was born from that frustration: a local-first, privacy-respecting study companion that lets you systematically build and verify your knowledge across every domain that top AI labs actually test on.
Why This Tool?
Motivation
- • Targeted Coverage: Curated questions from real AI/ML interview loops at top labs — not generic software engineering trivia
- • Privacy First: Runs 100% locally with no mandatory cloud dependencies — your study progress stays on your machine
- • Science-Backed Learning: Built around active recall and spaced repetition, proven to dramatically improve long-term retention
- • End-to-End Prep: Covers technical, behavioral, and negotiation stages — not just coding questions
Study Modes
- • Question Bank: Flashcard-style review with LaTeX math rendering and diagram support for deep theoretical questions
- • Quiz Mode: Self-assessment with optional AI-powered feedback via OpenAI or Anthropic APIs
- • Mock Interview: Timed practice sessions that simulate real interview pressure and pacing
- • Debug Scenarios: Hands-on ML training debugging exercises — a staple of research engineer interviews
Domain Coverage
- • LLM Post-Training: RLHF, DPO, reward modeling, and alignment fine-tuning
- • Agentic AI & Memory: Tool use, planning, retrieval-augmented generation, and continual learning
- • Reinforcement Learning: Policy gradients, value methods, multi-agent systems, and environment design
- • Safety & Evaluation: Model alignment, red-teaming, benchmark design, and robustness testing
Progress & Retention
- • Spaced Repetition: Automatically surfaces questions you're weakest on at optimal review intervals
- • Domain Mastery Tracking: Visual progress across all covered domains so you know exactly where to focus
- • Curated Resources: Linked reference materials for deeper study on any topic
My iOS App
iOS · Featured Project
A beautifully crafted iOS application built with SwiftUI, featuring intuitive design and smooth user experience. The app showcases modern iOS development practices with clean architecture and responsive animations.
Key Features
User Interface
- • SwiftUI Framework: Modern declarative UI framework for building native iOS interfaces
- • Custom Animations: Smooth transitions and micro-interactions for enhanced user experience
- • Responsive Design: Adaptive layouts that work seamlessly across different iPhone screen sizes
- • Dark Mode Support: Fully integrated with iOS system appearance settings
Technical Highlights
- • Clean Architecture: MVVM pattern with separation of concerns
- • Data Persistence: Core Data integration for local data storage
- • Networking: Async/await based API calls with error handling
- • Performance: Optimized rendering and memory management
