I build AI-native products, automation platforms, and marketplace systems in B2B SaaS — from 0→1 through enterprise scale.

What I ship:

• AI & Automation — LLM-powered knowledge systems, NLP retrieval, AI deflection pipelines. Built KB intelligence at Hiver that cut resolution time by 17% and drove 2× adoption across 20+ enterprise pilots.
• Marketplace Platforms — Two-sided liquidity design, API integrations, conversion optimization. Scaled Amber's marketplace to 180+ partners with a 4.6× conversion lift (5% → 23%).
• Workflow Engines — Event-driven automation, SCIM/API provisioning, multi-tenant enterprise infrastructure. Took Zluri's automation platform from ~$200K to ~$3M ARR.

How I work:
Signal first approach, I instrument everything, write real PRDs, partner closely with the customer, engineering, and design, and iterate on data, not intuition alone. I care about what moves the needle: activation, adoption, ARR.

Currently exploring: AI agent systems for ops/support, API-first micro-SaaS, and platform extensibility patterns.
Excited about: Voice AI — real-time speech interfaces, voice-native product design, and where conversational AI goes when it stops being text-first.

Hinglish Whisper Benchmark · openai/whisper #2761

Evaluated OpenAI's whisper-1 on 20 Indian-context audio clips across 6 categories — pure Hindi, Hinglish, Indian proper nouns, addresses with PIN codes, and call-center phrases.

Key finding: no output path exists for romanized Hinglish. Whisper switches the output script entirely — Latin-script input returns as Devanagari. Structural blocker for Indian support-call and WhatsApp transcription pipelines.

Dataset: huggingface.co/datasets/Primepluto/hinglish-whisper-benchmark

I write about product thinking, AI systems, and B2B growth on x https://x.com/primepluto_ and LinkedIn
Portfolio: rudransh.work
Email: Rudransh2k@gmail.com

Open to: Product leadership conversations, AI/automation collaborations, early-stage startup opportunities 📍 Bengaluru · LinkedIn · Resume