Fractional Tech & Product Advisor for SaaS & AI. Founder, Aarohii. I help teams ship AI features people can trust.

github

Rohann Niggam • Peerlist

An open-source checklist (MIT) for testing AI and LLM features before customers do. Five steps: define "good" as a number with an eval set, write a "never" list held to zero failures, run the evals in CI on every prompt and model change, put a budget on cost per request, and watch production with a rollback switch. Includes an example eval-cases.yaml you can copy and start with ten cases.

AI Feature Readiness Checklist

Small businesses run sales across WhatsApp, spreadsheets and separate payment apps, so nobody has the full picture of a customer. Captverse puts deals, payments and WhatsApp conversations on one customer record, with AI agents that help with the follow-ups. Built at Aarohii for teams that have outgrown spreadsheets but don't want a heavy CRM.

Captverse

ViraQueue writes social media drafts in your voice and schedules them across LinkedIn, Instagram, Facebook, X, Threads, YouTube and more. The difference from other AI social tools: AI never publishes on its own. Every post waits for a human to approve it, from the dashboard or from Telegram on your phone. Built for founders and small teams who want consistent posting without handing their brand to an unsupervised model.

ViraQueue

Functional tests confirm a button works. They don't notice it moved four pixels, lost its hover state, or disappeared behind a banner on one browser. PixellPeep catches those visual breaks before customers see them. Upload the screenshots you already capture or let PixellPeep capture URLs, and one CI manifest turns dozens of screens into a single pull-request check. Six comparison algorithms, from pixel-exact to perceptual, so you set strictness per screen. Live since June 2026.

PixellPeep

Architected comprehensive automation framework that reduced release cycle time by 40% across 5 product teams. Implemented using Python, Selenium, and Jenkins with CI/CD integration.