Built — applied work.

Ask the book anything — the assistant answers only from the contents of Digital Karma, citing the passages it drew from, so responses stay grounded in the text rather than the model's imagination. Follow-up suggestions keep the conversation moving, an optional SMS hand-off delivers highlights to your phone, and every exchange passes through curated admin review to keep quality high.

Tech stack

• Perplexity Agent API — reasoning & retrieval
• RAG grounding over the full book text
• Dependency-free vanilla JS embed widget
• Dedicated API service at book.himanshukalkar.ai

One intelligent layer over both Google and Outlook calendars. It handles smart scheduling and public booking links, builds prep panels before each meeting, blocks focus time around your energy patterns rather than just free slots, and pairs upcoming conversations with relevant ideas from Digital Karma via a curated content assistant.

Tech stack

• FastAPI backend
• PostgreSQL
• LLM-powered scheduling & content pairing
• Google Calendar + Microsoft Outlook integrations
• Runs at calendar.himanshukalkar.ai

Sweeps four variables — main-room and lock-off occupancy, and both nightly rates — through every step of their ranges to compute all 3,520 possible scenarios, deterministically. Explore the outcome space as an animated swarm plot, a yield histogram, an occupancy-by-rate heatmap, or live sliders that rank any scenario as a percentile against the full grid. Runs entirely in the browser; nothing is uploaded.

Tech stack

• React 18 (UMD build)
• Tailwind CSS
• Babel Standalone — in-browser JSX
• Single-file HTML, zero backend — fully client-side