Mrigesh Thakur

Coder , die-hard manga reader and student

[Mrigesh Thakur @mrigesh_thakur]

" What I cannot create, I do not understand " ~ Richard Phillips Feynman

Inspired by this, I'm embarking on a new journey: building Machine Learning algorithms from scratch.
My goal is to ditch the black-box libraries and truly understand the math behind the magic.

I'll be starting with Classical Machine Learning — the core ideas that built the foundation for modern AI.
Join me as I build and document everything in public.

GitHub: https://t.co/1jf5wQkOML
(star the repo, fork it — trust me, it'll be a ride)

#MachineLearning #AI #Python #DataScience #BuildInPublic

PS: I'll be tweeting my progress, learnings, and code breakdowns here day by day.

[Mrigesh Thakur @mrigesh_thakur]

I've always felt that cutting creative testing when CAC rises is the wrong move. Audience overlap seems like a much better lever to pull. But I'm curious , do those echoVME reviews ever talk about a third lever: optimizing what the landing page does with the traffic you've already paid for ?

[Mrigesh Thakur @mrigesh_thakur]

Cheap CPMs on the older-male side only pay if the advertorial and the page get rewritten for that reader if I am not wrong, since most of these offers inherit a page built for the 45+ women auction. Of the eleven, which do you feel has the widest gap between the ad and the page it lands on?

[Mrigesh Thakur @mrigesh_thakur]

@midomidas Do you have a number of days you wait before calling it, or is it more feel ?

[Mrigesh Thakur @mrigesh_thakur]

@IsmailLate52662 Do you get to ship them in stages on client work, or is it always one push?

[Mrigesh Thakur @mrigesh_thakur]

@oliviercroguy I totally agree here with you the format gets treated as the fix when it's really a longer runway for the same argument, so when you inherit an account doing this, how do you tell a bad advertorial from a bad offer ?

[Mrigesh Thakur @mrigesh_thakur]

@ConvertCraft @kaperbusiness @joshh7878 @replohq @EuNaturalHealth Quick question the comparison format you told about , does it work because it lets the reader disqualify the alternatives themselves rather than being told to or
Does it hold up when the competitor set is weak, or does it need real rivals to push against?

[Mrigesh Thakur @mrigesh_thakur]

@antonioventre_ The failure mode I see more often is the cap set too low, which quietly removes you from the auctions worth winning. Nothing breaks, volume just never arrives.

[Mrigesh Thakur @mrigesh_thakur]

@SEOKeval What was the tell for you? I have walked away for the same reason and I still cannot name the signal before the second call.

[Mrigesh Thakur @mrigesh_thakur]

@jordanfry The ops lead sending it unprompted is the tell. They feel the lead volume before anyone on the marketing side opens a dashboard.

[Mrigesh Thakur @mrigesh_thakur]

@val_shah @Shopify @taboola @Teads Does the gut health angle need a longer advertorial than the weight loss ones, or does that reader get to the offer just as fast?

[Mrigesh Thakur @mrigesh_thakur]

@JackHenryPaxton What I keep running into is a concept that holds CPA at $5k and quietly falls apart at triple the spend. Finding the winner is the easy half.

[Mrigesh Thakur @mrigesh_thakur]

Hi Afolayan, we’re building an autonomous product growth engineer that helps companies create hyper-personalised product experiences and continuously improve conversion through UX experimentation.

In an early case study, our core optimisation algorithm increased conversion by 60%. We’re currently looking to partner with design operators and agencies that want to offer measurable growth outcomes alongside their design work.

Our pricing is performance-based: $99 for every $5,000 in additional revenue generated. The product website goes live next week, and I’d be happy to send you early access along with a short Loom walkthrough of the case study.

https://t.co/edr9Oy4Nfy

Would you be open to a quick conversation?

[Mrigesh Thakur @mrigesh_thakur]

@RWebmasters Following the open build on this one. Curious how ClarityIQ will verify a fix actually held once BFCM traffic hits, holdback or before and after?

[Mrigesh Thakur @mrigesh_thakur]

Hey Azi , came across your product-page redesigns and really liked how you focus on changes that improve conversion, rather than just making the page look better.

We’re building Gradient to help CRO designers go from identifying friction to generating testable UX variants and running experiments much faster.

Quick demo:
https://t.co/VPxPcMjbPc
Website:
https://t.co/8ius4DdpB4

Given the number of eCommerce and CRO projects you’ve worked on, I’d genuinely love your take on whether this could fit into your client workflow—or potentially explore collaborating.

Open to a quick 15-minute call ?

[Mrigesh Thakur @mrigesh_thakur]

@Chinwe_Owanta Hey! We’re building Gradient to help eCommerce brands identify conversion issues, generate improved UX variants, and test what actually performs better.

Here’s a quick demo: https://t.co/bc91dXGkvR

Would love to connect: https://t.co/EJzxzdJdgZ

[Mrigesh Thakur @mrigesh_thakur]

This is a great use of Gemini for identifying CRO opportunities 👏

We’re also building Gradient to take the next step : That is analyze the full website, break pages into editable components, generate improved UX variants, and continuously test what actually increases conversion.

Quick demo:
https://t.co/VPxPcMjbPc
Website:
https://t.co/8ius4DdpB4

Would love to connect with eCommerce founders and CRO experts interested in trying it.

[Mrigesh Thakur @mrigesh_thakur]

Hi @hovinthenorth just came across your CRO work through this post and thought Gradient might be relevant to what you do.
https://t.co/8ius4DdpB4

We’re building an autonomous product growth engineer that analyzes websites, identifies conversion friction, generates UX variants, and helps test which changes perform better.

Here’s a short Loom: https://t.co/VPxPcMjbPc

Would you be open to a quick call?

[Mrigesh Thakur @mrigesh_thakur]

Hi @designkast just came across your CRO work through this post and thought Gradient might be relevant to what you do.
https://t.co/8ius4DcRLw

We’re building an autonomous product growth engineer that analyzes websites, identifies conversion friction, generates UX variants, and helps test which changes perform better.

Here’s a short Loom: https://t.co/VPxPcMiDZE

Would you be open to a quick call?

[Mrigesh Thakur @mrigesh_thakur]

@OpenAI , I was not expecting this feature to be this good 🤌

[Mrigesh Thakur @mrigesh_thakur]

Hi Fabian, we’re building an autonomous product growth engineer that helps companies create hyper-personalised product experiences and continuously improve conversion through UX experimentation.

In an early case study, our core optimisation algorithm increased conversion by 60%. We’re currently looking to partner with design operators and agencies that want to offer measurable growth outcomes alongside their design work.

Our pricing is performance-based: $99 for every $5,000 in additional revenue generated. The product website goes live next week, and I’d be happy to send you early access along with a short Loom walkthrough of the case study.

Would you be open to a quick conversation?