Project 01 — VisaFile
DS-160, minus the suffering.
The DS-160 can take hours of form-filling. VisaFile turns your answers into an automated application run, stopping only when it actually needs you.
Built by Animesh SharmaOPEN TO GOOD IDEAS & GOOD CONVERSATIONS
I figure things out.
Then I
build them.
Engineer, product person, automation obsessive, and professional “give it to Animesh, he’ll figure it out” person.
Ask me anythingBUILDSHIP
01 / Show, Don’t Tell
This is the part where I prove it.
Some came with the job. Others were an itch I had to scratch.
Context in,
reply out.
reply out.
Customer
Plan
Conversation
12° 14′CONCIERGE 01Project 02 — AI Insurance Concierge
Support that knows what’s going on.
An AI support system that understands the customer, their journey, the insurance plan, and the conversation before drafting a reply.
Product and engineering at Gradly · Animesh SharmaProject 03 — Gradly Links
The useful kind of short story.
An internal Rebrandly alternative with a custom Gradly domain—built in-house to manage branded short links and save the company about $400 a month.
Built at Gradly by Animesh SharmaProject 04 — AI Claims AdjudicationUpload the bill.
Upload the bill.
We’ll handle the rest.
Upload a medical bill and bank details. The app checks eligibility, sends the reimbursement by ACH, and confirms it by email—payment lands in 1–2 business days.
Product and engineering at Gradly · Animesh SharmaExperiments
Side projects that started as an itch and became real products.
Sortify
Rediscover the songs you already love.
Reads the musical traits of your saved songs to spot patterns and draft playlists. You review, tweak, and approve every one before it reaches Spotify.
Crate
Find what you saved for a reason.
Turns an Instagram export into a private, searchable library. It finds related captions, tags, and notes on your device, then suggests collections you can shape yourself.
You’ve seen the work. Ask for the story.
Ask about a project02 / About
Engineer on paper. Problem solver in practice.
I’m at my best when the problem is messy.
Incomplete requirements. Bad documentation. Too many people involved. Nobody quite knows the answer. And somehow, it still needs to ship Friday.
Good.
Give me half the context and I’ll find the other half. I’ll read the code, talk to the user, ask questions, and pull at threads until the problem makes sense.
That mentality took me beyond engineering — into product, operations, customer experience, partnerships, and strategy.
I kept picking up problems until my job title had to catch up.
The titles changed.
The job didn’t.
Figure it out.
Software just happens to be my favorite form of leverage.
If it’s repetitive, automate it. If the process is broken, rebuild it. If everyone’s solving the symptom, find the actual problem.
And if nobody quite knows how to do that yet?
Now you have my attention.
03 / Experience
I kept picking up problems until my job title had to catch up.
Now
Chief of Staff
Gradly Inc.
The problems got bigger. So did the scope.
I work across strategy, product, engineering, partnerships, and operations — leading a 10-person cross-functional team while staying close to the product.
2023 — 25
Technical Lead
Gradly Inc.
I started building products. Then I started building the team.
Owned products from problem to production across AI, claims, payments, mobile, and carrier integrations.
Grew engineering from 1 to 5–6 people, interviewed 30+, hired 8, and integrated 5 insurance carriers.
2021 — 23
Founding Engineer
Gradly Inc.
A few scripts turned into the foundation.
Joined as the only engineer and went from automating manual work to building the backend, internal systems, insurance integrations, and an early AI prototype.
That foundation grew into a platform serving 10,000+ members across 25+ U.S. universities.
I was building the machinery it ran on.
04 / Right now
Currently building, learning & following whatever looks useful.
- 01Automating work that has no business being manual
- 02AI that works outside the demo
- 03Building for founders at Hire Animesh
05 / Notes
Opinions, loosely organized.
Things I’ve learned from building software, breaking software, fixing operations, talking to users, and occasionally doing things the hard way.
01
Context is something you acquire.
“I don’t have enough context” is useful for about five minutes.
4 min read02Prototype before architecture.
Walk ten steps and check the map before walking a kilometre in the wrong direction.
5 min read03I don’t automate tasks. I automate roles.
The interesting part of automation starts when you stop thinking in individual tasks.
6 min readAllThe rest of the notebook.
More thoughts on building, learning, and figuring things out.
Explore the blog