Growth @Rifa AI | Making AI work like your most reliable teammate | Previously @ Topmate | 30u30 IQ

AI isn’t replacing people. It’s replacing processes that slow them down.

At Rifa AI, we’re building intelligent agents that automate back-office conversations across chat, voice, email, and support without breaking ops.

The goal isn’t to sound human.
The goal is to get work done, with AI that can reason, act, and complete entire workflows across tools.

Our best users don’t even realize they’re talking to a bot.
That’s the point.

Before this, I led growth at Topmate and picked up a few hard truths:
 → Shiny doesn’t scale
 → Funnels lie
 → Speed breaks things, but it also builds them

Rifa is a tougher game. The stakes are higher.
You’re shipping into systems where mistakes cost trust, not just clicks.

That’s why I care about:
 → AI that’s functional, not performative
 → Growth that’s compound, not chaotic
 → Shipping like it’s a reflex, not an event

If you’re building at the intersection of AI, ops, and velocity, let’s chat.

Shubham Khoker

I just saw Ankit's post about Topmate hitting profitability, and I couldn’t scroll past without smiling.

To the world, "profitability" is a financial metric. To me, having been in the trenches with Ankit and Dinesh, I know it’s a badge of pure grit.

I’ve said this before, and I’ll say it again: My time as Head of Growth at topmate.io was my real MBA.

It was the most intense classroom I’ve ever stepped into. Working with both of you gave me my earliest instincts for product, growth, and just... how to operate in absolute chaos without losing clarity.

A lot of the DNA I carry today as a founder comes straight from those months. The pace, the difficult decisions, and the relentless execution.

You don’t forget teams like that.

Seeing them hit profitability is a reminder of what consistent execution looks like when most people would have given up.

Huge congratulations to Ankit, Dinesh, and the entire team.

This win is well-earned. So proud to have been a small part of this journey.

Everyone’s busy calling themselves “AI-first.”
But your customers don’t care about it. 

When we first started pitching to healthcare clients, I thought leading with “AI-powered” would make their eyes light up. Instead, it usually made them lean back in their chairs, because for most healthcare stakeholders, AI-first sounds risky.

They’re not thinking about the model you used.
They’re thinking: what happens if it fails? Will it violate compliance? Leak data? Or give the wrong answer to a patient?

So now, I start those conversations differently.
About what we’ve done to make sure AI actually works when it’s supposed to.

Every voice agent we build passes through a simple checklist before going live:
→ Does it log every interaction for audit trails?
→ Does it know when to stop and escalate to a human?
→ Does it meet every compliance protocol, even if no one’s watching?

Once that foundation is clear, then the conversation about automation begins.

AI might get you in the door, but risk management keeps you in the room.
That’s what actually builds long-term trust.

I predict AI will vanish by 2030.
And I say that as someone who spends every single day building with it.

Not literally, but in how we talk about it. 
Five years ago, AI was a buzzword.
Today, it’s a feature.
Five years from now, it’ll just be invisible infrastructure.

No one will brag about “AI-first” anymore, just like no one says “internet-first” today. Because the best technology always disappears behind the experience.

When we deploy voice agents in hospitals, no patient ever says, “Wow, this AI bot is amazing.”

They just say,
“Finally… someone actually picked up.”
That’s the direction we’re headed. AI won’t live in headlines or product launches.

It’ll live inside workflows, quietly fixing things that humans don’t have the time or patience to.

So, here’s what I’m betting on:
𝟭/ 𝗔𝗜 𝘄𝗶𝗹𝗹 𝘀𝘁𝗼𝗽 𝘀𝗼𝘂𝗻𝗱𝗶𝗻𝗴 𝘀𝗺𝗮𝗿𝘁 𝗮𝗻𝗱 𝘀𝘁𝗮𝗿𝘁 𝗯𝗲𝗶𝗻𝗴 𝗱𝗲𝗽𝗲𝗻𝗱𝗮𝗯𝗹𝗲.
We’ve hit peak hype, and the next winners won’t be the flashiest demos; they’ll be the ones that never break in production. 

𝟮/ 𝗥𝗲𝗴𝘂𝗹𝗮𝘁𝗶𝗼𝗻 𝘄𝗶𝗹𝗹 𝗯𝗲𝗰𝗼𝗺𝗲 𝗮 𝗺𝗼𝗮𝘁.
In healthcare, reliability is the foundation. The companies that bake guardrails in from day one will outlast the ones chasing speed.

𝟯/ 𝗜𝗻𝘁𝗲𝗿𝗳𝗮𝗰𝗲𝘀 𝘄𝗶𝗹𝗹 𝗱𝗶𝘀𝗮𝗽𝗽𝗲𝗮𝗿.
Typing, clicking, searching; all of it will fade into natural conversations. Your voice will become the universal remote for how you access systems.

And maybe that’s the irony:
AI will be everywhere, precisely because it will feel like it’s nowhere.
By 2030, the best AI won’t feel like AI at all, it’ll just feel human.

If that happens, what’s one part of your workflow you hope disappears first?

Building should feel hard.
But it shouldn’t feel blurry.

Most people assume burnout happens when you work too much.
In my experience, it’s when you don’t know what you’re working towards anymore.

The chaos creeps in quietly: 
→ Too many priorities
→ Too many opinions
→ Too many half-baked “let’s just test it” plans

Suddenly your calendar’s full, your brain’s fried, and you're still not moving.
And the worst part?
You start confusing activity with progress.

I’m not anti-hustle.
I’m anti-chaos.

The only thing that’s ever pulled me out of a spiral wasn’t rest—it was clarity.
Clarity of: 
→ What problem actually needs solving
→ What to drop (even if it’s shiny)
→ What not to chase this quarter

Because burnout isn’t just about energy.
It’s about direction.

And sometimes the most productive thing you can do is… stop and think.
That’s not a “growth hack.”
That’s survival.

Tag a founder who needs to hear this.

The first time I had to let someone go, I didn’t sleep that night.

I kept replaying the conversation in my head; the silence after I said their name, the way they nodded slowly, the awkward gratitude at the end.

Nothing prepares you for that moment.
When we hired our first person at Rifa AI, it felt like validation, like someone believed in what we were building.

But when it doesn’t work out, it hits differently.
You start asking yourself if you could’ve coached better, communicated sooner, or just seen it coming.

That’s when I learned something that never makes it into startup books:
Company culture isn’t defined by how you celebrate wins. It’s defined by how you handle endings.

Hiring is about hope.
Firing, when done right, is about honesty.
Both take a piece of you.

And both, in their own way, make you a better builder — and maybe, a better human.

When you’re a non-technical founder building a technical product, impostor syndrome moves in next door, rent-free.


There have been times when the team was deep in conversation about inference speeds and vector databases, and I’d be there, nodding, jotting notes, then Googling half the terms later, only to realize I still didn’t get all of it.

However, I’ve learned that it’s part of the job.
I’m lucky to have a co-founder who lives and breathes the tech side (that’s Sameer). My role has always been to connect what we’re building to what the customer actually feels.

So I stopped pretending to understand everything and started asking the simplest questions again.

Every time I asked a “dumb” question, the conversation got clearer for everyone in the room. We stopped over-explaining to each other and started building something users could actually use.

Over time, it has become my strength. I stopped trying to sound technical and focused on what I can understand and build with.

Yeah, impostor syndrome still drops by sometimes, but I’ve learned to make peace with it.

If you’re a non-technical founder quietly sitting through a technical meeting, feeling like an outsider, you’re not behind.

You’re just seeing the problem through a different lens.
And sometimes, that’s exactly what your team needs.

Everyone says “talk to your customers.”
I’m taking that advice one notch higher, and I’ll pay you for your time.
No clickbait. Stay with me till the end.

As we explore the healthcare vertical at Rifa AI, I’ve realised something important:
You can’t automate an industry you don’t understand.
You can’t sit behind dashboards and guess how frontline workflows operate.

So before we ship anything in healthcare, I want to learn from the people who actually hold the system together. The ones who deal with chaos, patient load, insurance calls, and operational bottlenecks every single day.

That’s why I want to speak with:
→ Front desk staff at hospitals/clinics
→ Patient coordinators
→ Medical receptionists
→ Nurses who work closely with ops/patient flow

I’m not looking for confidential data. I’m looking for operational reality. I want to know where the friction is, what the "unwritten rules" of your workflow are, and what tools drive you crazy.

And yes, I’m happy to compensate for your time.
It could be a paid virtual meeting or a conversation over a coffee.

If you work in US healthcare operations and are open to a 30-minute paid chat (or know someone who fits), drop me a DM, tag them below, or share this post.

If you’re planning to build a startup, here’s the part that no MBA or mentor ever tells you.

Last week, when Anisha Jain and the HIVE team at the University of Chicago asked me to speak to their students, I said yes instantly.

Because I remembered what it felt like to be in their shoes.
Hungry to build something, unsure where to start, and confused by all the startup terms around “AI,” “product-market fit,” “fundraising,” and “the right time.”

Back then, I thought progress would be about finding the perfect idea or learning the right framework. But, once you get past the excitement and the theory, you hit the part where nothing looks linear, certainty disappears, and you have to build even when your confidence doesn’t show up with you.

That’s the phase that actually teaches you what matters, and here’s what I learned during that phase:
1. Start with a problem that keeps bothering you.
You don’t need to know how to code on Day 1. What you really need is a problem you can’t ignore... something you’re willing to think about even when no one’s watching.

2. A co-founder is someone who believes in the same future you see.
Skills can be hired, but what you can’t hire is someone who shows up on hard days with the same energy as you do.

3. Investors look for clarity more than big pitches.
If you can explain what you're building, who it helps, and why now is the right moment, the conversation changes instantly.

4. Your early job will teach you more about building a company than any “startup guide.”
For me, most of my instincts came from Topmate, the messy days, the mistakes, the adrenaline, the long nights. That’s where I learned how startups actually work.

None of this shows up in the pitch decks or the headline interviews. But it’s exactly what determines whether you ship something real or just talk about it.

So, if you’re at that stage: excited, overwhelmed, curious, questioning everything, that’s exactly where most real journeys begin.

P.S. What’s one lesson your founder journey has taught you?

I thought I knew what leadership was.
Then I became a founder.

Turns out, leading one branch of the company and leading a company are two completely different things.

When I was working as a Growth Head, it was about ownership within a structure.
We had a clear direction, a set goal, and a team to rally.

Someone had already built the walls of the house; my job was to optimize the rooms, fix the wiring, and ensure growth. There was stability, even in chaos.

At Rifa, I quickly realized leadership from the ground up is a whole different game.

There is no existing structure. You are literally building the walls while simultaneously deciding what kind of house you even want to live in.

The shift is profound, and honestly, sometimes scary.
You stop being a functional head and become the emotional backbone of the entire operation. Every product pivot, every client bet, every tough hiring call has effects that you simply cannot delegate away.
 
They need someone who can admit, "Yes, this is fog," and still be the first one to take a step forward.

That's the real leadership muscle I had to build, and that’s the part most leadership books skip:

The silence that comes with being responsible for everything, and the humility it takes to show up every day, even when you don’t have all the answers.

If you've made the jump from employee to founder, what surprised you most about leadership?

People keep asking me, "When will your AI be smart enough to replace a nurse?"

I tell them that's the wrong question.

The right question is, "How can our AI be reliable enough to handle the 5 hours of administrative work that burns a nurse out every single day?"

The goal isn't to replace our heroes, it's to give them room to be human. Isn’t it?

You don’t need a dev team or a $50K budget to build a voice agent anymore.
Hard to believe, but that’s where we are now.

I’ve seen companies spend months wiring together APIs, vendors, and compliance checks just to make a phone line “talk.” By the time it goes live, the use case has already changed.

That’s the gap we’re closing with Rifa.
Building automation shouldn’t feel like building another product; it should feel like setting one up.

With our platform, you can design, test, and deploy a compliant AI voice agent in minutes. 
You don’t need to know coding. 
You can simply drag, drop, and go live on any phone number.

And it doesn’t stop when the call ends.

→ Your CRM updates automatically.
→ Follow-ups get scheduled.
→ Issues get flagged without anyone touching a dashboard.

Because real automation shouldn’t add another layer of complexity, it should remove it. That’s what we’re betting on at Rifa: making AI powerful and practical for the people who actually use it.

If building automation were this simple, what’s the first workflow you’d hand off?

We are building an AI Voice Agent to automate phone calls. But to get our first 10 customers, I had to make hundreds of phone calls myself.

That’s the irony of building in AI today. 
Everyone wants to automate everything, including product development and GTM.

But here’s the reality I’ve seen time and again, from leading growth to now building Rifa…
AI can automate workflows, but can’t automate trust.

And when you are selling into industries where a single mistake has consequences— healthcare, finance, compliance-heavy ops—that difference matters.

These buyers aren't looking for a "cool AI tool." 
They're worried because the risk surface has expanded:

_ "What if the response is wrong?"
– “Who is liable for the model’s error?”
– “How do I explain this to an auditor?”
– “What does this break in my existing workflow?”
– “What happens if the model behaves unpredictably?”

These aren’t objections you solve with a funnel, or a case study, or an automated sequence.
These are objections that need context, interpretation, and reassurance. 

And this is where old-school GTM suddenly becomes a superpower again because it helps buyers process risk at a human pace.

That’s why the AI companies that are winning right now have strong trust-building mechanisms:
• High-context conversations
• Non-scalable onboarding
• Long-form demos
• Domain fluency
• Founder-led selling in v1
• Field intelligence from real calls

Because GTM in AI is about risk translation before efficiency or growth. 

If your product is AI-first, your early GTM must be risk-first. 
And right now, risk is still interpreted by humans, not workflows.

See, AI will eventually sell AI.
We’ll get there.

But we're in the messy middle, where technology has leapfrogged, but buyer psychology hasn’t caught up.

The biggest mistake you’re making as a founder is trying to make your GTM move as fast as your tech. While in reality, it should match the pace of your buyer’s fear.

Don't let the hype trick you into hiding behind your screen.

P.S. Are you also seeing traditional channels outperform the automated ones right now?

Machine Learning @ Call of Duty | Rebel Data Scientist | Click follow to join a community of data, gaming, and tech professionals 🎮

Databricks

As a Field CTO in the Databricks gaming vertical, I am an external technical strategist helping game studios unlock the full potential of their data. 

As a hands-on solutions architect, I design and implement scalable, future-ready systems that empower developers to build next-generation player experiences, run efficient operations, and turn innovation into impact.

My role blends technical strategy, architectural design, and industry evangelism. I work directly with game studio leadership to architect the infrastructure that enables smarter game design and better live operations.

Ultimately, I help studios create leverage with their data so they can focus on what matters most: building great games that players love.

Field CTO, Gaming

ggAI

We are the next generation of future-ready game data geeks - engineers, analysts, ML practitioners, and technical leaders - united by a belief that collaboration and openness are the keys to unlocking better player experiences. 

Whether you're scaling infrastructure, tuning personalization models, or wrangling event data, GGAI is where the gaming industry's sharpest minds come together to share best practices, solve hard problems, and build what’s next.

Founder

Rebel Data Science

Rebel Data Science brings a refreshingly authentic approach to educating both people and brands working on AI. By making data science and AI more accessible, we hope to open up doors for women to find fulfilling careers in data. We are excited about partnering with brands who share our commitment to our communities.

Microsoft

Director, Engineering

Activision

As a Director for Call of Duty, I led a team focused on game security. 

We utilized a suite of technology to make decisions in real-time and strengthened the technology that delivers low-latency concurrent experiences to hundreds of millions of players.

Senior Manager, Machine Learning Engineering

As an expert machine learning engineer, I implement and oversee machine learning projects to detect and minimize disruptive behavior in online gaming.

Staff Machine Learning Engineer

As a Senior Data Scientist on Call of Duty’s behavioral machine learning team, I worked at the intersection of big data engineering, real-time data analytics, and behavioral modeling to drive business value in both paid and F2P online games. 

I balanced hands-on model development while collaborating cross-functionally with my stakeholders across Activision.

•	Managed a team of data scientists developing distributed machine learning models to support behavioral machine learning initiatives
•	Built, refined, and executed a roadmap for a behavioral machine learning approach to anti-disruptive behavior
•	Outlined team best-practices including library packaging, versioning, containerization, CI/CD and task-scheduling
•	Triaged and shaped incoming requests and supervise team deliverables

Senior Data Scientist

Game Security Group (GSG)

The Game Security Group (GSG), is the video game industry's preeminent information-sharing and analysis center that servers as a central resource for exchanging information on cyber threats. The GSG is comprised of 23 companies and facilitates in-person and virtual meetings for members to discuss security challenges and approaches to defending their corporate and game infrastructure.

Governing Council Member

Well Data Labs, Inc

As a Data Scientist on the Well Data Labs Research and Development team, I focused on how data science and machine learning can impact operational efficiencies in the oilfield and increase the ROI of unconventional wells. 

•	Developed, deployed and patented static and real-time machine learning models to automate the recognition of events in time-series metered data
•	Conducted customer discovery and authored studies that addressed key market needs while proposing solutions and strategies to business challenges
•	Prepared and delivered internal and external industry presentations to effectively demonstrate the value of data science in oil and gas
•	Worked collaboratively in an agile environment with Product and Engineering teams on prototyping, designing and implementing robust production-ready business solutions
•	Created documentation and project artifacts for maintaining production code for machine learning platform
•	Explored and visualized data to find actionable insights and developed robust pipelines to automate reporting for Professional Services

Data Scientist - R&D

Associate Data Scientist

AirDNA

As a Product Analyst at AirDNA, I leveraged, visualized and transformed short-term rental data into strategic, actionable analytics to inform decisions resulting in 120% YOY growth and a projected $6.8 million run rate.
 
•	Acted as the point-person on all things related to user data and analytics, from data ingestion, maintenance and tracking, to understanding and reporting
•	Prepared and delivered executive presentations to strategically meet growth objectives
•	Explored and visualized data to find actionable insights and make product recommendations through funnels, churn risk analyses, cohort analyses, customer lifetime value, and behavioral segments
•	Utilized a comprehensive understanding of the full-stack framework to implement and integrate front-end and back-end analytics solutions 
•	 Created documentation and project artifacts such as analytics schema, user stories, process diagrams, work-flows, and technical requirements
•	Prepared specialized market analyses for content creation in high-level publications such as the New York Times and Wall Street Journal

Product Analyst

Analytic Partners

As a Marketing Analytics Consultant at Analytic Partners, I built predictive models to improve marketing ROI and executed communication strategies across five divisions at two Fortune 100 companies. 

•	Developed predictive attribution and media mixed models
•	Outlined the impact of advertising for five individual divisions across two Fortune 100 companies
•	Addressed client specific business challenges while executing communication strategies and presentations to convey actionable insights
•	Extracted and synthesized information from multiple data sources while proactively performing data validation to ensure accuracy, quality and integrity of data
•	Work collaboratively as a consultant with colleagues in sales, marketing, and finance to provide data-driven recommendations for client resource allocation

Marketing Analyst

University of Colorado

As a Computational Researcher, I worked at the intersection of theoretical physics, chemistry, and computational modeling to simulate the behavior of electrons in photochemical reactions.

•	Utilized quantum simulation packages on massively parallel super-computing systems to model the behavior of electrons in photochemical reactions
•	Explored systems which had previously been limited to a mixed QM/MM approach
•	Author of 3 published articles, one of which was in a Top 10 journal
•	Earned grant worth over 1 million computational hours on the 8th largest supercomputer in the world
•	Worked with large, complex data sets generated by massively parallel super-computing systems
•	Spearheaded successful and ongoing summer internship program for at-risk high school students focusing on computer science, chemistry and biotechnology

Computational Chemist

- All views shared are my own opinion and do not reflect the stance of my employer

🎰 Vegas born
🧪 Ex-chemist, current data scientist, lifelong gamer. 
👩‍🔬 Advocate for women in STEM, tech, and gaming.
🚨 Security strategist at Activision. 

Have you ever wondered “what comes after breaking into data science?” I share tips on how to excel in your career - advocate for yourself to get promoted, navigate situations with leadership, and mindfully leverage your strengths to optimize your career. 

Carly is a data scientist, computational chemist and machine learning engineer. She obtained her M.S. in chemistry from the University of Colorado focusing on computational quantum dynamics. She has authored multiple peer-reviewed publications and holds two non-provisional machine learning patents. 

When she isn't writing about herself in the third person, building mechanical keyboards or neglecting the oxford comma, she works as a public speaker, author and content creator focusing on AI. 

More specifically, she applies Python-based machine learning and best practices in statistical analysis to creatively solve novel problems. This includes harnessing various toolkits to model real-world structured and unstructured data and drive business value to stakeholders at all levels of the organization. 

Areas of Expertise: data science, machine learning, fraud detection, trust & safety, distributed systems, leadership, team-building, time-series and dynamic modeling, anomaly detection, biotechnology, high-frequency signal processing

Languages:
- Python (NumPy, Pandas, PySpark)
- SQL
- Scala

ML Platforms:
- TensorFlow
- Keras
- Scikit-learn

Engineering:
- AWS
- GCP
- Docker
- Apache Spark
- Flask
- Airflow
- CI/CD (Jenkins, Kubernetes)

Carly Taylor, M.Sc.

If you want to read a happy Thanksgiving post, just scroll right on by this one…

Things are not okay right now for a lot of people. 

I’m sure you’ve noticed too - viral posts about fellow professionals who are about to lose everything. They’ve been laid off, unable to find work, and some are about to be homeless. 

I don’t want this to become the new normal. 

Being so lost and hopeless that you are begging for work on LinkedIn is, in fact, not normal. 

And it flies in the face of everything you’re told about people who struggle.

“Oh they don’t have the right degree.” 

“They must be addicts.” 

Nope. Just regular people like you and me. 

The astounding wealth transfer in the U.S. driven by layoffs, AI, and tied up with a pretty little bow we call “shareholder value” is not normal. 

There is an armed guard at my local grocery store checking receipts because of theft. And instead of addressing issues like inflation and price gouging by the grocery store, we accept this as normal. 

We accept that the cashier who checked us out is on food stamps. 

Corporate greed at this magnitude is not normal. 

I have a lot to be thankful for. I could spout some happy platitudes as I sit here in the airport writing this note. 

But that’s not why I’m here. And I don’t think it’s why you’re here either. 

I am using my platform to say none of this is normal or okay.

And I’m tired of it.

taplio

The Battle

Shubham Khoker

Carly Taylor, M.Sc.

Followers

Posts

Engagements

Likes

Performance ScoreShubham Khoker's LinkedIn Performance

Performance ScoreCarly Taylor, M.Sc.'s LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Shubham Khoker

Carly Taylor, M.Sc.

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Shubham Khoker's LinkedIn Performance

Performance ScoreCarly Taylor, M.Sc.'s LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreShubham Khoker's LinkedIn Performance