Om worked across AI, analytics, and execution: building chatbot workflows, creating dashboards and reports, and helping different teams turn ideas into usable systems.
He shipped Sakhi AI, supported the RF × Anemia careplan through launch, explored messaging platforms, helped teams take data-driven decisions, and ran AI agents in parallel. Most of all, he learnt and grew here at Saathealth with people across teams.
Here are Om's 360° Feedback responses. Six colleagues shared how they experienced working with him. The hexagon shows the numbers, and the cards highlight what stood out.
Saathealth's first production AI app. Hospicash sales-training chatbot, live at sakhi‑ai‑beta.vercel.app. Stack: Next.js + TypeScript, Gemini 2.5 Flash for reasoning and multi-modal understanding, Sarvam AI Bulbul v3 for Indian-language STT + TTS. Demoed at the Feb 2026 all-hands.
Supported the 12-week WhatsApp careplan through to completion. My role was managing and unblocking, working closely with the Saathealth and REAN Foundation teams to see it across the line. The clinical engagement numbers (88.76% Week-1 engagement, 66.67% Week-12 retention across 67 enrolled women) belong to that team; I helped the program ship and the dashboard reach them.
Recorded marketplace and product data, then presented it in weekly and monthly reviews: week-on-week movement, retention snapshots, first- and second-order cuts of the underlying datasets to surface patterns worth discussing. I was the reporter, not the decider.
Ran a longitudinal journey for 150–200 users across different products over a 12-week window. The discovery was about engagement rate, where people actually stay vs. drop off. That changed how we drafted outbound: bold / italics for scannable emphasis, a tighter emoji system, and the rule of one key message per image.
For OmniVec-ICMR, built a sample-data dashboard for the campaign. This project is still ongoing and I'll keep contributing to it as it develops. On the language side, built the Saathealth Language Translation chatbot as a Gemini Gem trained on how Varun frames outbound content: the tone, humour and positioning, so translations across 7 Indian languages feel like Saathealth rather than just a translation.
"A lot of improvement needed in communication and clarity of thoughts, especially to someone without tech understanding."
"Step back and think neutrally before you execute."
"With more clarity in thoughts and understanding of the larger picture, Om will do wonders."
"Lead with the destination: state your conclusion or main recommendation first, and then back it up with supporting data if asked. Don't take the listener on the journey of how you got there; just give them the destination."
"The Rule of 3: before you speak, pause and force yourself to group your thoughts into a maximum of three bullet points. If you have ten points, find a way to group them into three themes."
"The energy and initiative are genuinely there. What needs to match is the clarity on what problem is being solved before getting into building the solution."
"The shift to work on is from sharing the process to selling the conclusion. Output can sometimes feel more like experimentation notes than a clear, distilled recommendation."
"Solve your boss's problem and your boss's boss's problem. And be able to sell the simplest of the simple thing."
"Understanding what is needed at the hour and just doing it. Things need to get done. That is something you should build on."
"Write more. It helps you understand your thought process better. Start writing for yourself, it gives you clarity of thought."
Fair. Curiosity 4.5 and Ownership 4.0 got the work shipped. The 2.0 decides how far the work travels after it ships. That's the gap I'm taking into the MBA: tighter writing, data storytelling, and presenting to non-technical rooms. Lead with the destination, not the journey.
Here are the best memories from the last two years.







