Data Scientist and recovering high-energy astrophysicist. I also enjoy running, if somewhat slowly. All views are my own.
Southampton, UK
Joined November 2008
- Tweets6.8K
- Following569
- Followers581
- Likes1.9K
Pinned Tweet
The 2017 Total Solar Eclipse as seen from Swan Valley, Idaho from first to last contact. #cottonwoodeclipse #eclipse2017 #astrophotography – at Idaho, USA
I'm running the @CardiffHalf Marathon next month and will be supporting Cancer Research UK. Make a donation to @CR_UK and support defeating Cancer cardiffhalf26.enthuse.com/pf…
Get 15% off your first Overherd Oat Milk Powder order: overherd.uk/products/oat-mil…
Adam Hill retweeted
i just dropped a full walkthrough on how i built and sold an AI agent for $5,000 in 24 hours
inside, i break down:
- the exact problem it solved and why they paid fast
- tools + setup used to build it
- how I packaged and closed the deal in one day
reply "500" + like + repost and i’ll DM you the full breakdown.
(must be following)
Honestly, the Google Pixel 10 Pro has been brilliant — camera, battery, everything just works. If you’re tempted to switch, here’s 15% off with my referral code 👇
REF-SOEDWQ8Y7K41MOX0PR1ODVL
store.google.com
Adam Hill retweeted
Built ai ad agents with:
sora2. nano banana. higgsfield. veo3.
what agencies charge $20k+ for…
and take 6 weeks to deliver…
now takes minutes and costs cents
last month those agents pulled $122k
this month we’re scaling toward $830k/month
200+ ads daily.
flawless avatars.
zero agency fees.
rt + comment "122k" and i’ll dm you the system
(must follow for dm)
Adam Hill retweeted
Canva + laptop + internet + 60 mins/day = $369 per day.
I created a step-by-step guide (normally $77).
Free for the next 24 hrs.
To get it:
• Follow me (so I can DM)
• Like & Repost
• Reply "Canva"
Who doesn't need a new bank account that rewards you for joining?! join.monzo.com/c/yqzgfdrj
Final stage of the MLOpsZoomcamp is to mark and review other projects: 1st one done; some good stuff there but not all aspects of the code runs :( and only managed to get partial debugging
🚀 Yesterday I finished my student grade prediction project exploring how to build a full end-to-end MLOps showcase! From env management to monitoring, here’s what I learned:
This thread isn’t just about grades, it’s about building a complete, robust ML system: tracking experiments, orchestrating workflows, serving models, monitoring outputs, maintaining code quality, and automating deployment.