All work
FiloProduct Manager (Founding US PM)Aug 2022 to Aug 2023

A tutoring promise built for the US.

A successful marketplace playbook did not automatically fit a new market. I used student research to shape US-specific matching around collaborative learning and a clear promise on response time.

Trust mechanism: a 60-second matching promise, kept for 98% of US requests in under 45.

US users
120K
In six months
ARR
$1.5M
US marketplace, in six months
of US requests
98%
Matched in under 45 seconds
activation
+35%
After funnel improvements
From request to live learning
  1. Upload problem
  2. OCR intake
  3. US-specific matching
  4. Tutor accepts
  5. Live session
98%of US requests matched
in under 45 seconds
0s45s60s target
The marker shows the measured matching result against the target. Individual step durations are not implied.

My scope

I owned the US launch across matching, activation, OCR intake, and the session experience. I worked with engineering, data science, and operations on student and tutor workflows.

The Problem

A working playbook still needed a local test.

Filo had 2M+ users when I joined. The existing marketplace provided a starting point, but bringing it to the US required understanding whether its matching experience met local student expectations.

What I Found

Students wanted to learn together.

Across 100+ user interviews, collaborative learning emerged as a key need for the US experience. That called for more than translating the interface. Matching, onboarding, and the session itself needed to support that expectation.

The decision

Chose
US-specific matching for collaborative learning.
Over
Port the existing India matching model.
Evidence
100+ user interviews informed the US experience.
Trade-off
Adapt the matching and session experience instead of reusing the existing playbook unchanged.
Cost
Dedicated work on US matching, onboarding, and session continuity.
What Shipped

Make the promise visible in the request path.

The experience connected uploaded problems, OCR intake, US-specific matching, tutor acceptance, and a live session. Matching had a 60-second target. Session continuity and low-bandwidth audio/video supported the experience after the match.

I worked across the activation funnel and student and tutor onboarding, with quality thresholds and fraud detection supporting marketplace trust.

System Details

Intake
Upload a problem and use OCR to support intake.
Matching
US-specific logic with a 60-second target.
Measured result
98% of US requests matched in under 45 seconds.
Sessions
Tutor acceptance, session continuity, and low-bandwidth audio/video.
Trust
Tutor onboarding, quality thresholds, and fraud detection.
Research
100+ user interviews.
Outcomes

A US marketplace with a promise it could keep.

The US marketplace went on to reach 120K users and $1.5M ARR in six months. 98% of US requests matched in under 45 seconds against a 60-second target.

Activation increased 35% and conversion reached 3%+.

What I Learned

Local insight belongs in the core product.

Market entry is a product decision, not just a distribution decision. Trust came from adapting the experience to how students wanted to learn and then delivering on a measurable promise.

Next case study · ProductSquads

Making unfamiliar documents reliable.

Read case study
Let’s connect

Let’s talk product.

For senior PM and AI product opportunities, or a complex problem worth working on.

hello@duasahil.com
LinkedInRésumé