Work With Me

Book the call. Decide After.

20 minutes, no obligation. We'll look at what you're trying to solve, whether I'm the right person to solve it, and what the next step is if the answer turns out to be yes.

Atharva
Atharva Badhe Data Scientist & Freelancer

Prefer email? Reach me at hello@atharvabadhe.com

Or send a brief

Rather start async? Fill this in and I'll come prepared with ideas for your specific situation.

THE CALL ITSELF

What we'll talk about in 20 minutes

The call has a structure. Five things, in this order:

1
The business question this work is meant to answer.

One sentence is fine. We refine it together.

2
Where your data lives today.

A rough sketch is fine. Stripe, your app database, a spreadsheet someone in ops maintains, an HR tool nobody remembers buying. All of it.

3
What you've tried before, and what didn't hold up.

Failed prototypes welcome. They tell me more than success stories.

4
Whether the work you have in mind is a dashboard, a model, an automation, or something else.

Most founders walk in with the wrong category. We figure that out together on the call.

5
Whether I'm the right person for it, and if not, who is.

If we're not a fit, the call ends with a name or two of someone who is.

By minute eighteen, you have a clear answer on whether to move ahead, and I have enough to send a written proposal within 48 hours.

If we run out of time on one of the five, the rest moves to a written follow-up. The call stays at 20 minutes.

COME AS YOU ARE

What you don't need to bring

A lot of founders postpone the call because they think they need to have it all worked out first. They don't.

Bring nothing if you have nothing. The call still works.

Here are the four most common reasons founders delay booking, and the actual answer to each:

"I haven't decided what I want." That is the call. Half the founders I talk to come in asking for one thing and leave with a different recommendation. The scoping call is where the right answer gets named.

"My data is a mess." Every founder I've worked with has said this. The first week of every engagement is a data audit. Mess is the baseline, not an exception.

"I'm not sure if this is even an ML problem." Often it isn't. A SQL query and a Slack alert solve half the problems that get framed as model problems. I'll say so on the call if that's where we land.

"It's too early for a project like this." Possibly. We'll know in twenty minutes. If you are too early, I will say so and point you toward a free path that helps you get to the right moment faster.

The call costs nothing. The proposal that follows it costs nothing either. The first cost in the whole process is the engagement itself.

AFTER YOU HANG UP

What happens after the call ends

The call ends. Here is what comes next.

Within 48 hours, a one-page proposal lands in your inbox. It includes scope, deliverables, timeline, a fixed quote, and the two earliest start dates I can offer. Plain English. No surprise terms.

If we're a fit but the timing isn't right, the proposal includes a hold-the-slot option. You can lock in a start date up to eight weeks out without committing to the full quote. The hold costs nothing. You can release it any time before the start date.

If we're not a fit, the proposal becomes a written referral. One or two names of freelancers or small agencies I would send my own clients to, with a one-line note on why each is better suited for what you described.

You are free to take whatever time you need with the proposal, or to leave it unanswered. The call is a conversation. Nothing on this site behaves like a funnel.

A SMALLER FIRST STEP

Start with a one-week paid pilot

A four-to-twelve-week engagement is a lot to commit to before you know how I work. The pilot is the smaller first step.

The pilot is a focused one-week task that produces a real artifact you can use even if we never work together again.

Common pilots:

  1. A data audit: A written report of where your data lives, what is joinable, what is missing, what is broken, plus a one-page recommendation on what to fix first.
  2. A single dashboard: One chart or one report that answers one specific question you have been answering manually.
  3. A feasibility study: A written go-or-no-go on whether the ML or RAG idea you have is worth building, with the evidence behind the call.
  4. A model baseline: A scikit-learn or LightGBM baseline on your data so we know the floor before any decision on a production build.

The pilot is paid. Pilot pricing is sized to fit a one-week scope and stays well below my standard project rate. If the fit is right, the pilot rolls into a full engagement and the pilot fee is credited toward the larger quote. Otherwise the pilot artifact and the code are yours to keep.

The pilot still includes the scoping document, the Friday log, the Loom walk-through, and the same async cadence as a full engagement. You see the working rhythm before the larger commitment.