AI Product Builder | AI Voice Agents for Hospitality | Full-Stack Developer

Bengaluru, Karnataka, India

Bengaluru, Karnataka, India







Manipal University Jaipur

buildspace
Join Shagun and 10k+ other professionals on
Tal is the app where bosses meet engineers & builders directly, without middlemen
Download Tal BossAI Product Builder | AI Voice Agents for Hospitality | Full-Stack Developer
I'm a recent Information Technology graduate with a strong interest in product management, AI, automation, and software development. I enjoy working at the intersection of technology and business—understanding user problems, validating ideas, and building solutions that create real value. Over the years, I've worked on projects ranging from web applications and recommendation systems to AI-powered tools and automation workflows. Currently, I'm building Mira (https://mira-landing-alpha.vercel.app/) , an AI voice agent for hospitality businesses that helps handle guest inquiries and reservations. Through this experience, I've been learning firsthand about customer discovery, product strategy, user research, go-to-market, and the challenges of taking a product from idea to MVP. I'm always excited to connect with builders, operators, and people working on interesting problems. Whether it's discussing product ideas, AI, startups, or emerging technologies, I'm constantly looking to learn and grow.
Mumbai
claude code
claude
elevenlabs
huggingface
Pipecat
Vercel AI SDK
Associate Product Manager at YourStory Media, Aug 2026 - Present · 2 mos, Bengaluru, Karnataka, India
AI Research and Developer Intern at Laneway, Dec 2025 - Mar 2026 · 4 mos, Bengaluru, Karnataka, India, Built a full-stack meeting intelligence platform — from Chrome extension to production backend. - Engineered a Manifest V3 Chrome extension capturing Google Meet sessions with real-time per-attendee analytics (camera, mic, speaking events, join/leave timestamps); published to the Chrome Web Store - Designed a 5-stage AI processing pipeline: MD5 deduplication → Redis cache → Faster-Whisper transcription → HuggingFace pyannote.audio speaker diarization → multi-LLM task extraction with fallback routing across OpenRouter, Gemini, OpenAI, and Ollama - Architected a RAG-based assistant using pgvector for semantic retrieval over tasks, decisions, and meeting history - Implemented dual-auth backend (JWT + API key), 4-state meeting lifecycle, Google Calendar & Gmail OAuth 2.0 integration; deployed on Render with Cloudflare R2, D1, and Workers
Manipal University Jaipur
buildspace
github: https://github.com/shagunverma-04