How Claude Help Me Build My $200k+ ML Resume

, I want to breakdown exactly how I used Claude to help me craft my resume that landed me a $200k+ Machine Learning Engineer offer.

Let’s get into it!

Resume Mindset

Your resume is the single most important document in your career. I spent over 10 hours drafting mine. That’s a lot of time, but it’s exactly why it got results.

Something I tell all my coaching clients:

Resumes are more of an art than a science

This is not to sound “smart”, but rather to tell you that there is no “definitive” guide on what your resume needs to look like.

I will give you general principles and guidelines for what a good resume looks like in this video and how you can optimise it with Claude.

However, a great resume is perfect for the job you are applying for.

Tailoring your resume is the ultimate cheat code and the easiest way to tailor a resume is through having a “master” resume that has everything on it.

Don’t worry about the length, how many bullet points you have or the number of technologies you list.

Then, you create a resume tailored to the job description and role by removing unnecessary or irrelevant information.

It’s much easier to subtract than add, and it’s a much quicker method to tailor your resume to every application.

Fundamental Structure

Before letting Claude go wild on your resume, you need a solid underlying structure.

We need rough sections and raw content that give the AI enough to work with so it actually “gets” the tone, achievements, and overall vibe you’re going for.

I’ve tried generating a resume from scratch with Claude, and it doesn’t work well. You still need to feed it your technical skills, experience, and education for it to draft something decent.

In my experience, it’s far more effective to start with a rough draft using a clean template, then have Claude optimise it into a top 1% resume.

Let’s break it down section by section, so you know exactly what to include.

Design & Layout

People overcomplicate resumes so much. Your resume needs to be clean, simple, and, to some extent, boring!

No fancy formatting, colours, and definitely no multi-columns.

All we need is black text, consistent readable fonts and bullet points.

If you want a proven template that has landed me and my clients $200k+ job offers, then I recommend checking my template linked below:

Data Science & ML Resume Template
_Get the exact resume that landed me $200k+ data science & ML offers_resume.egorhowell.com

Structure

Structure can vary, and you should tailor it to each role. The goal is to get the most important information showcasing your fit for the job into the top third of the page.

However, in general the following works well for the majority of cases.

• Header
• Summary Statement
• Skills
• Experience
• Projects
• Education

You only really need the first 4 sections in the majority of cases.

Header

All you need is:

• Your name.
• Job title or how you see yourself. (Put data scientist if in doubt, don’t say “aspiring”).
• Contact details.
• Location.
• Relevant links like LinkedIn, GitHub, Medium, Kaggle etc. No more than 4.

One thing to ensure is that the links work.

Summary Statement

A summary section used correctly can instantly tell the recruiter that you are fit for the job; it can also be a detriment if you spew rubbish and fill it with too much exaggeration.

You realistically should only have a summary section after 2+ years of experience in the field, as it allows you to “summarise” your expertise and showcase the results you have generated.

There is a slight exception, though: when you are changing roles. This is to provide the recruiter with context for your application based on your previous experience.

• Years of experience.
• Demonstrate results.
• Specialisms.
• No fluffy language such as “passionate,” “determined,” or “hard-working”
• No cliches or vague language.

There is no “correct” summary as it will really depend on your situation and the role you are applying for.

Technical Skills

For data science jobs, technical skills are so important and you will immediately be filtered on these, hence why you should have them right at the top of your resume.

You should only have three rows for this section

• Languages — Split by Proficient and Familiar:
• Proficient — Programming languages you feel comfortable working in.
• Familiar — Programming languages you have worked in the past with, but not overly confident at the moment.
• Technologies — List tools you have worked with or feel comfortable working in e.g. AWS, Git, Docker, GCP, Linux, Bash.

Experience

This is the most essential part of your resume, but arguably the most challenging to get right.

Start by drafting your experience bullet points and then we will refine them with Claude to make them really standout.

However, the better your drafts are, the better Claude will optimise them, so start with the following:

• 3–5 bullet points max experience.
• Don’t use your actual job title; you are allowed to be “creative” and make it sound more impressive, so it stands out.
• Only have experience that is relevant for the job.
• Weight your most recent experience more, or the experience most relevant to the role.
• Use active action verbs all the time, like “developed,” “led,” “optimised,” “generated,” and “improved” as it’s clear you did the work.
• Weave in technical details, don’t be too scared to be too granular and stuff in keywords.
• Use numbers, metrics and especially financial impact in every bullet point. If you struggle with this, it’s where Claude will come in handy and I will show you how to do this later on.

Projects

Projects are very similar to experience, but it’s often hard to measure business impact because the work isn’t for the business.

Either way, you need to think of them as experience because that’s the closest “experience” you have to the job role if you are adding this section in.

Here are some general principles to follow:

• Make sure to use numbers and metrics, as you would for the experience, in every bullet point.
• Provide links to your projects, and make sure they actually work.
• Make sure your project’s GitHub or portfolio website is well laid out, with a clear README.
• If you can add unit tests, a front end dashboard or website and any other end-to-end technologies you used to built the project.
• If you have many projects, add the most relevant ones for the role you are applying for

Education

This section doesn’t need to be too complicated, all you need is:

• University and degree
• Grade or expected (if its good, if not leave it blank)
• Graduation or expected graduation date

Following these checklists will give you a sufficiently good resume that we can now use Claude to optimise even further.

Claude Optimisations

I want to go over the three most impactful optimisations Claude can do for your resume.

Financial Impact

Many of my clients tell me how hard it is to link their work to financial impact.

To be blunt about it: if you work for a company, everything you do ties back to financial impact in some way. They employ you to generate a return on your salary.

So there’s always a way to connect your work to the bottom line; you just need to get creative with it.

Enter Claude.

You might not immediately see how your work connects to financial impact, but Claude can help guide you through that process. 

I want you to use this prompt:

“[Paste your bullet or task description here]”

Walk me through this step by step:

Ask me clarifying questions about what this task/project actually did — the scale, who used it, what process it touched.

Identify the possible financial levers this work could plausibly connect to (e.g. cost savings, revenue growth, time saved, risk/error reduction, retention, efficiency gains).

For each lever you identify, tell me exactly what data or numbers I’d need to find to quantify it (e.g. “hours saved per week x hourly cost of that role” or “% reduction in errors x average cost per error”).

Tell me where I might realistically find that data (my own project notes, team dashboards, a manager, finance/ops team, etc.).

Once I bring back the numbers, help me rewrite the bullet in an “action → method → quantified result” format.

Do not invent or estimate any numbers on my behalf — only help me identify what to look for and how to calculate it once I have the real data.

And let’s say I give it this bullet point from my resume:

Produced a suite of CatBoost models to predict risk prices. This involved building an automated modelling pipeline in Databricks using MLflow with optimised training using Bayesian Hyperparameter Tuning (Hyperopt).

I will skip all the boring back and fourth details you will need go into, but a rewrite of this bullet point may look like this:

Built an automated CatBoost risk pricing pipeline in Databricks (MLflow, Bayesian hyperparameter tuning), reducing model retraining time from [X hours] to [Y hours] per cycle and improving pricing accuracy by [X]%, contributing to a [X]% improvement in loss ratio.

Tailoring & Removing

Most people have used Claude to tailor their resume to certain roles, but I actually don’t like this because it ends up making things up.

What’s better is simply to ask it to remove irrelevant information from your master resume to make it only a page long.

This is incredibly simple, and to be honest you can do this yourself, but I found the AI will be very ruthless compared to myself.

ATS & Job Description Alignment

This is a unique one, and it’s not simply asking Claude to stuff keywords, but rather re-write existing terms to be more aligned.

You want to use it to compare your resume’s language against the job description’s language, and flag where you’re using different words for the same skill (e.g., you wrote “forecasting model,” the job posting says “predictive analytics”).

Many ATS systems and recruiters scan for exact-match terms.

Here is a sample job description excerpt (Senior Data Scientist, Retail/E-commerce):

We’re looking for a Senior Data Scientist to lead predictive analytics initiatives across our supply chain. You’ll build machine learning pipelines for demand forecasting, work with MLOps best practices, and communicate business impact to stakeholders. Experience with cloud platforms (AWS/GCP/Azure) and hyperparameter optimisation is essential.

My bullet point was:

“Implemented a new recipe popularity forecast machine learning model using LightGBM that improved the lead day 5 forecast by 33%. Algorithm was deployed on AWS through lambdas and step functions.”

Suggested rewrite from Claude to close the gaps:

“Built a predictive analytics model for recipe demand forecasting using LightGBM, improving 5-day-ahead forecast accuracy by 33% and generating £500,000 in annual food waste savings. Deployed on AWS via Lambda and Step Functions as part of an automated MLOps pipeline.”

These are very subtle changes, but improving all these little things by 1% compounds and makes you standout massively.

Implementing everything I have mentioned in this article from the fundamentals to optimising with Claude will genuinely make your resume in the top 1%.

However, a good resume alone will not get a $200k+ machine learning job, you also need to master networking and all the various interviews.

If you want to speed run that process, then I recommend you apply to code to careers where we can help you do exactly that!

Apply here!

See you soon!

Connect With Me

• YouTube
• LinkedIn
• Instagram
• Website