Aaron Hays

Authority Signal

Founder

Alterfy

Co-Founder & Director of CS

Rockbot

Account Director

Optimizely

Manager - Onboarding / Implementation

ZoomInfo

Customer Onboarding  Manager

Gordon Food Service

Project Manager

Booking.com

Account Manager

I help SaaS leaders build trust, visibility, and inbound opportunity on LinkedIn | Writer on SaaS leadership, reputation building, and content that converts

I'm currently in a full-time SaaS leadership role while building Authority Signals, where I help other SaaS professionals grow their presence, sharpen their voice, and attract opportunity, without needing to go viral.

After more than a decade in SaaS, leading CS teams, building onboarding playbooks, and navigating the emotional pressure of being “the glue” - I realized something:

Great work doesn’t speak for itself.

Many of the most capable operators I know stay invisible online. They lead quietly, burn out privately, and get passed over for louder, flashier people.

I help them change that.

I work with SaaS leaders - especially in CS, GTM, and RevOps, to clarify their voice, tell the truth about their leadership journey, and build real visibility on LinkedIn.

I’m not a coach, and I’m not here to make you post every day.

Through my Leadership Narrative Sprint, I help mid-to-senior professionals reposition themselves around trust, clarity, and authority, not hustle or fluff.

I also write about the emotional reality of modern leadership:

Burnout from invisible labor

Boundaries in high-stakes roles

Building trust without performative branding

✅ 16K+ followers
✅ 1K+ newsletter subscribers
✅ Posts seen by over 1 million people

If you’re tired of doing great work in silence, and you want your next opportunity, client, or promotion to find you, I’d love to help.

📩 DM me “Leadership” to see how the Sprint works.
Or follow for weekly insights on trust, visibility, and voice.

🤖Data Scientist | Turning Raw Data into Money Generating Insights | Data-led Growth for Your Business 🚀

Freelance Data Scientist | ML Engineer | SQL | Python | Power BI | MLOps

Data analytics and Machine learning models built 5X faster - no quality compromise. Worked with 14 clients in 7 different industries in the last 3 years. Helped current company scale to $5M+ ARR. Featured TOP RATED PLUS talent on Upwork in 2025 with 100% job completion and satisfaction.

My Toolkit:
- Data analytics: Python, SQL, Excel
- ML Engineering: FastAPI, AWS, Streamlit, great_expectations
- Data visualization: PowerPoint, PowerBI, Plotly, Matplotlib
- Data modeling: Scikit-learn, TensorFlow, Keras
- MLOPs: CI/CD (GitHub Actions), Docker
- Additional tools: Notion, n8n, Canva, Git/GitHub

My Stats:
- 5-stars reviews on 99% of projects delivered.
- 100% Job Success Score for on-time, budget delivery.
- Helped scale currently business to $5M ARR with ATS analysis & Reporting.
- 1.5 years TOP RATER PLUS on Upwork.
- 1800 hours billed.
- $60k+ earned.

What You’ll Get Working With Me:
- A data analyst who is a real problem solver.
- Quick responses (within the hour on working hours)
- Clean, maintainable code and detailed documentation
- A clear “NO” if something isn’t feasible; no wasted time
- A well-crafted report or dashboard as per the requirements

My clients think:
- "Anas is great and highly knowledgeable. He is passionate about solving data science challenges and has helped me significantly streamline some of my processes, enabling better scalability.
- "Anas, was prompt and thorough and I would def use him again in the future."
- "Anas is great to work with. He made the job so fluid and his knowledge of data science makes him someone worth listening to. I’m glad I chose him for my small business. I’m looking forward to what’s coming next, and I would definitely hire him again."

Popular Clients & Products:
- Uhubs: Financial company based in London
- AllAboutGroup: Law recruitment firm based in London
- Growth Dynamics: Marketing and data strategist based in London.
- Adway: Tech and recruitment company based in Gotheborg, Sweden

About me:
- Football fan since I can remember (Real Madrid fan)
- Tennis lover and practitioner (one-handed backhand style)
- Mangas fan (One Piece, Naruto, Bleach)
- Travel addict (South Asia, Europe, Africa, America (soon))
- Reading is something I enjoy as much as coffee (oat latte fan recently)

-----

YouTube Channel
→ GO HERE: https://www.youtube.com/anas_riad

Request services via LinkedIn or Email below:
r.riadanas@gmail.com

Hire on Upwork:
https://www.upwork.com/freelancers/~0163377bf3029c7553

Anas Riad

Not every problem needs an LLM. 
Some need 12 characters of regex.

I see people spending $𝟬.𝟬𝟯 𝗽𝗲𝗿 𝗔𝗣𝗜 𝗰𝗮𝗹𝗹 to extract emails from text.

Meanwhile: ⤵️

[𝗮-𝘇𝗔-𝗭𝟬-𝟵+_.-]+@[\𝘄-]+\.\𝘄+

Done. 𝗭𝗲𝗿𝗼 𝗰𝗼𝘀𝘁. Runs in milliseconds.

We're in 2026 and somehow we forgot that basic code still works. 

Not everything needs a prompt, a model, and a 3-second response time.

Need to validate a phone number? 
→ Regex. 

Need to clean whitespace from 10,000 rows? 
→ A one-liner. 

Need to rename 500 files? 
→ A bash script your laptop runs in 0.2 seconds.

LLMs are incredible for reasoning, writing, and ambiguity. 
But when the task has a clear rule and a predictable pattern, just write the rule.

Cheaper. Faster. More reliable.
Stop using a rocket ship to cross the street.

PS: Do you use LLMs for everything because it's convenient?

♻️ Repost if your network needs this reminder 
📌 Save it for the next time you almost open ChatGPT/Claude for string matching

Follow me for more → Anas Riad

Data Scientist & BI Consultant @Adway • ML Engineer • YouTube @ Anas Riad

The ML project you wish you built sooner.
2.5hr of hands-on learning ⤵️

No excuses for skipping end-to-end ML now! 🤣

(PS: Link to the full YT video in the comments)

When I first started working on churn prediction, 
I noticed something that always bothered me.

There are tons of Regression projects out there, 
but 𝟗𝟗% 𝐨𝐟 𝐭𝐡𝐞𝐦 stop at the notebook.

It’s like drawing the blueprint for a house but never building it.

So I created a full end-to-end regression project that goes from:
raw dataset → trained model → deployment → tests → cloud.

Here’s what I cover in this project & video:
→ Split data by date (time-series aware)
→ Data Quality with Great Expectations
→ Well-structured pipelines (feature, training & inference)
→ Unit & smoke tests to keep things reliable
→ Experiment tracking with MLflow
→ Containerization with Docker
→ FastAPI serving layer
→ CI/CD with GitHub Actions
→ AWS (ECS, ECR, S3, ALB) for Deployment

It’s the same type of stack I use in freelance + personal work, 
but applied to regression.

Looking forward to your thoughts on it.

🌟 Link to the full YT video in the comments

PS: I have another end-to-end coming soon, stay tuned 🤔

YouTube is still the BEST place to learn ML.

Most people still watch the wrong videos.

Here are 𝟭𝟮 𝗬𝗼𝘂𝗧𝘂𝗯𝗲 𝘃𝗶𝗱𝗲𝗼𝘀 𝘁𝗼 𝗺𝗮𝘀𝘁𝗲𝗿 ML ⤵️ 

---

𝗦𝘁𝗮𝗻𝗳𝗼𝗿𝗱 𝗖𝗦𝟮𝟮𝟵: 𝗠𝗮𝗰𝗵𝗶𝗻𝗲 𝗟𝗲𝗮𝗿𝗻𝗶𝗻𝗴
 
The legendary starting point for ML.
Covers core theory, intuition, and key algorithms.

Video: https://lnkd.in/e-aJRQun
 
---

𝗔𝗹𝗹 𝗠𝗮𝗰𝗵𝗶𝗻𝗲 𝗟𝗲𝗮𝗿𝗻𝗶𝗻𝗴 𝗖𝗼𝗻𝗰𝗲𝗽𝘁𝘀 𝗘𝘅𝗽𝗹𝗮𝗶𝗻𝗲𝗱 𝗶𝗻 𝟮𝟮 𝗠𝗶𝗻𝘂𝘁𝗲𝘀

A fast, structured overview of all major ML concepts.
Great for building a mental map before going deeper.
 
Video: https://lnkd.in/exCP58F6
 
---

𝗠𝗟𝗢𝗽𝘀 𝗖𝗼𝘂𝗿𝘀𝗲 𝗳𝗼𝗿 𝗕𝗲𝗴𝗶𝗻𝗻𝗲𝗿𝘀
 
Shows how ML actually works in real products.
Pipelines, deployment, monitoring, and production mindset.
 
Video: https://lnkd.in/erzM3b4F

---

𝗛𝗼𝘄 𝘁𝗼 𝗗𝗲𝗽𝗹𝗼𝘆 𝗠𝗟 𝗦𝗼𝗹𝘂𝘁𝗶𝗼𝗻𝘀 (𝗙𝗮𝘀𝘁𝗔𝗣𝗜, 𝗗𝗼𝗰𝗸𝗲𝗿, 𝗔𝗪𝗦)
 
From trained model to live service.
A practical deployment walkthrough with modern tools.
 
Video: https://lnkd.in/e3JtcAZk

---

𝗦𝘁𝗲𝗮𝗹 𝗠𝘆 𝗠𝗮𝗰𝗵𝗶𝗻𝗲 𝗟𝗲𝗮𝗿𝗻𝗶𝗻𝗴 𝗣𝗿𝗼𝗷𝗲𝗰𝘁 𝗧𝗲𝗺𝗽𝗹𝗮𝘁𝗲

Learn how to structure ML projects properly.
Clean folders, reproducibility, and real-world practices.
 
Video: https://lnkd.in/eds2Buq8

---

𝗘𝗻𝗱-𝘁𝗼-𝗘𝗻𝗱 𝗠𝗟 𝗣𝗿𝗼𝗷𝗲𝗰𝘁𝘀 𝗜𝗺𝗽𝗹𝗲𝗺𝗲𝗻𝘁𝗮𝘁𝗶𝗼𝗻
 
A full ML project from start to finish.
Data prep, modeling, evaluation, and delivery.
 
Video: https://lnkd.in/ebx_dvtn

---

𝗠𝗟𝗢𝗽𝘀 𝗣𝗶𝗽𝗲𝗹𝗶𝗻𝗲 𝗧𝘂𝘁𝗼𝗿𝗶𝗮𝗹 (𝗬𝗼𝘂𝗧𝘂𝗯𝗲 𝗩𝗶𝗲𝘄𝗲𝗿 𝗦𝗲𝗻𝘁𝗶𝗺𝗲𝗻𝘁)
 
Build a real ML pipeline using real data.
Feature engineering, modeling, and orchestration.
 
Video: https://lnkd.in/e9_xVb-Z

---

𝗬𝗼𝘂 𝗪𝗶𝗹𝗹 𝗡𝗢𝗧 𝗟𝗲𝗮𝗿𝗻 𝗠𝗟 𝗶𝗻 𝟯 𝗠𝗼𝗻𝘁𝗵𝘀
 
A reality check on learning machine learning.
Focuses on process, depth, and consistency.
 
Video: https://lnkd.in/eRqYhJWK

---

𝗘𝗻𝗱-𝘁𝗼-𝗘𝗻𝗱 𝗥𝗲𝗴𝗿𝗲𝘀𝘀𝗶𝗼𝗻 𝗠𝗟 𝗣𝗿𝗼𝗷𝗲𝗰𝘁
 
A practical regression project walkthrough.
Covers evaluation, assumptions, trade-offs.
 
Video: https://lnkd.in/e2GEnb4N

---

𝗘𝘃𝗲𝗿𝘆 𝗠𝗮𝗰𝗵𝗶𝗻𝗲 𝗟𝗲𝗮𝗿𝗻𝗶𝗻𝗴 𝗠𝗼𝗱𝗲𝗹 𝗘𝘅𝗽𝗹𝗮𝗶𝗻𝗲𝗱
 
A clear overview of common ML models.
 
Video: https://lnkd.in/eMW3YNKq

---

𝗣𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝗼𝗻-𝗚𝗿𝗮𝗱𝗲 𝗠𝗟 𝗣𝗿𝗼𝗷𝗲𝗰𝘁 𝗧𝘂𝘁𝗼𝗿𝗶𝗮𝗹
 
Focus on building ML systems that last.
Code quality, deployment, and maintainability.
 
Video: https://lnkd.in/eXYhdxgt

---

𝗜𝗻𝘁𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝗼𝗻 𝘁𝗼 𝗠𝗮𝗰𝗵𝗶𝗻𝗲 𝗟𝗲𝗮𝗿𝗻𝗶𝗻𝗴 (𝗙𝗼𝘂𝗻𝗱𝗮𝘁𝗶𝗼𝗻𝘀)
 
A simple introduction to ML definitions and concepts.
 
Video: https://lnkd.in/eT3qrGBt
 
---

📌 Save this for your next project.
♻️ Repost this to help your network in data.

90% of analysts use only 10% of Pandas.

Master these 10 operations. Level up instantly.

When I started learning data analysis,
I thought Pandas was “too big” to master.

Turns out… you only need the core.

Here are the operations I use almost daily:

• 𝗟𝗼𝗮𝗱 𝗱𝗮𝘁𝗮 → pd.read_csv()
• 𝗜𝗻𝘀𝗽𝗲𝗰𝘁 𝗾𝘂𝗶𝗰𝗸𝗹𝘆 → .head(), .info(), .describe()
• 𝗦𝗲𝗹𝗲𝗰𝘁 𝗰𝗼𝗹𝘂𝗺𝗻𝘀 → df[["col1","col2"]]
• 𝗙𝗶𝗹𝘁𝗲𝗿 𝗿𝗼𝘄𝘀 → df[df["col"] > value]
• 𝗚𝗿𝗼𝘂𝗽 & 𝗮𝗴𝗴𝗿𝗲𝗴𝗮𝘁𝗲 → .groupby().agg()
• 𝗖𝗿𝗲𝗮𝘁𝗲 𝗳𝗲𝗮𝘁𝘂𝗿𝗲𝘀 → df["new"] = expression
• 𝗠𝗲𝗿𝗴𝗲 𝗱𝗮𝘁𝗮𝘀𝗲𝘁𝘀 → pd.merge()
• 𝗛𝗮𝗻𝗱𝗹𝗲 𝗺𝗶𝘀𝘀𝗶𝗻𝗴 𝘃𝗮𝗹𝘂𝗲𝘀 → .isna(), .fillna()
• 𝗦𝗼𝗿𝘁 𝗿𝗲𝘀𝘂𝗹𝘁𝘀 → .sort_values()
• 𝗘𝘅𝗽𝗼𝗿𝘁 𝗰𝗹𝗲𝗮𝗻 𝗱𝗮𝘁𝗮 → .to_csv() or .to_excel()

That’s it.

If you can combine these properly, you can solve 80% of real-world problems.

The rest is practice and business thinking.

PS: Is Pandas your go-to library for data analytics?

📌 Follow Anas Riad for more content around data science

95% of data scientists who reach out to me

Have this 1 thing in common ⤵️ 

“They excel at building ML models in a notebook”
but "They struggle to build something beyond a Demo"

Look, I’m not against notebooks, but that mindset is restrictive.

If your project only:

→ runs on your machine
→ uses static data (all clean)
→ isn’t deployed and usable

Then it’s a "SIMPLE" demo, not a "REAL" data project.

In my recent journey to bridge the gap, I’ve decided to collaborate with Markus Kuehnle on a detailed carousel that includes major points of an end-to-end ML project, such as:

→ Setting project environment
→ Exploratory data analysis
→ Feature Engineering
→ Model Registry & Experiment Tracking
→ Feature, training, inference pipelines
→ Serving layer (FastAPI)
→ Docker container (reproducibility)
→ GitHub Actions (CI/CD)
→ Deployment & monitoring

Hope it gives you more clarity on what it takes to build and deploy an ML model end-to-end.

📌 Save this for your next project.
♻️ Repost this to help your network in data.

Machine learning End-to-End pipeline

I started in data analytics in 2021. 
The job barely looks the same today.

Back then, your value was tied to tools.

You learned SQL. You learned Python. Maybe some Tableau or Power BI. And that technical gap between you and everyone else in the company? That was your edge. You could dig deeper than the person pulling numbers from a quick Excel sheet. That was enough.

Fast forward to 2025 and 2026.

AI writes code faster than most junior analysts. It reasons through data problems. It spots patterns. And yeah, I know what some of you are thinking, "but it makes mistakes." Sure. So do we. We miss things. We overlook patterns. We carry bias into every analysis without realising it. We're not the gold standard we like to think we are.

I've been working with AI daily for the past 3 years now. And what I've watched happen slowly at first, then all at once, is that the technical barrier that made analysts valuable is shrinking fast.

That directly hits entry level roles the hardest.

If your entire value is "I can write a SQL query and build a dashboard," you're competing with something that does it in seconds. Not in a hypothetical future. Right now.

But here's what I've also noticed in 3 years of doing this work:

The analysts who actually move the needle? They were never just "tool operators."

They're the ones who sit with stakeholders and ask the hard questions. They're the ones who take a messy data output and turn it into a recommendation that helps the business earn more, save more, or move faster. They understand context. They push back on bad assumptions. They connect the numbers to actual decisions.

No AI tool does that well yet.

So if you're early in your career or thinking about getting into analytics the path forward isn't about memorising SQL syntax or chasing every new tool.

It's about:

→ Going deeper than surface-level reporting  
→ Learning to frame problems before jumping into data  
→ Giving recommendations, not just dashboards  
→ Expanding into predictive models, time series, and forecasting  
→ Using AI as a multiplier, not pretending it doesn't exist

The role of data analyst isn't dying.

But the version of it that was just querying and reporting? That chapter is closing.

The analysts who evolve will be fine. Better than fine, actually, because AI gives you leverage that didn't exist 3 years ago.

The ones who don't evolve will wonder what happened.

♻️ Repost if you think someone in your network needs to hear this  
📌 Save for when you need the reminder yourself

Follow me → Anas Riad

taplio

The Battle

Aaron Hays

Anas Riad

Followers

Posts

Engagements

Likes

Performance ScoreAnas Riad's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Aaron Hays

Anas Riad

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Anas Riad's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreAnas Riad's LinkedIn Performance