81k | Director, Data & Analytics/ AI @ Gartner

Summary

Making state-of-the-art Artificial Intelligence concepts accessible and profitable to professionals from any background for over 35 years!

Highlights
-  Veteran innovator with 3 decades of experience leading state-of-the-art AI and Analytics ventures
-  98% of students surveyed would recommend me to peers
-  Generative AI/ Agentic AI expert-  ChatGPT, Claude, Copilot, Perplexity, Lovable, n8n, more
-  Experienced AI instructor for Cisco, LinkedIn Learning, University of Pennsylvania, & Simplilearn
-  Senior AI Curriculum Advisor for Graduate School of Education at Ivy League university UPenn
-  Winner, International Society of Data Scientists Outstanding Voice Award
-  Social media leader with over 107k LinkedIn followers


Selected Clients
 
-  3M
-  Audi
-  Biogen 
-  Booz Allen Hamilton
-  Bristol-Myers Squibb
-  Cisco
-  Citicorp
-  Franklin Templeton
-  Gartner/ Evanta
-  Harvard Innovation Labs’ Experfy
-  IBM
-  International Olympic Committee
-  Los Angeles Times
-  Mercedes
-  Morgan Stanley
-  Royal Bank of Canada
-  St. Jude’s Children’s Hospital
-  Wells Fargo
-  US Homeland Security
-  US Veterans Administration 
-  Yale University

Partial Skills Inventory
 
-  AIOps
-  Accountable AI/ Ethics
-  Anomaly Detection
-  Anthropic/ Claude
-  Artificial Intelligence
-  AWS
-  Brokerage Firms
-  Client Management
-  Consulting
-  Compelling Presentations
-  Compliance Regs and Frameworks
-  Curriculum Development
-  Dashboards
-  Data Storytelling
-  Data Visualization
-  Education
-  Explainable AI
-  Feature Engineering
-  Financial Services
-  Fraud Detection
-  Gap Analysis
-  Generative AI
-  GitHub Models
-  Google Cloud
-  Google Gemini
-  Insurance
-  KPI's and Metrics
-  Lovable AI
-  Machine Learning
-  Microsoft Azure
-  MLOps
-  Model Ensembles
-  Nano Banana
-  OpenAI/ ChatGPT
-  Orange Data Mining
-  Perplexity
-  Postgres SQL
-  Professional Development
-  Research
-  Retail
-  Retention and Attrition
-  SPSS
-  SQL Server
-  Underwriting

T. Scott Clendaniel

🔬 #ArtificialIntelligence 𝗛𝗶𝗿𝗶𝗻𝗴- 𝗝𝘂𝘀𝘁 𝗪𝗵𝗮𝘁 𝗔𝗥𝗘 𝗘𝗺𝗽𝗹𝗼𝘆𝗲𝗿𝘀 𝗟𝗼𝗼𝗸𝗶𝗻𝗴 𝗙𝗼𝗿?

The Terrible Facts:

- Artificial Intelligence needs a lot more experts.
- Tons of qualified professionals aren't getting hired.
- The expectations of hiring managers are just out of control.

If you're a hiring manager, please STOP looking for an entire department in just one employee.  It simply doesn't work that way.  You need a team.

Personally, I find the job search to be truly dreadful. It is exhausting, soul-crushing, depressing, dispiriting- actually, I will run out time before I run of negative emotions on this topic!  

But I hold out hope that one day everyone else will figure this out!

Please follow me for more analysis! 

- T. Scott Clendaniel

🥳 || #AI Training for 3.1 Million Worldwide || 110K Followers || Follow me for genuine ROI from Artificial Intelligence and LinkedIn!

🔬 #AI for the Environment

The Ocean Cleanup provides remarkable technology in the form of the Interceptor 006, which I discovered courtesy of Pascal BORNET!

▶▶ Interceptor 006, the new Interceptor Barricade, shows early promise following deployment in Rio Las Vacas, Guatemala.

▶▶ Estimated 856,973 kg of trash removed and prevented from polluting the Caribbean Sea during installation and first trials; still awaiting anticipated ‘trash tsunami’ conditions to fully validate solution.

▶▶ Strong initial performance provides hope of ending annual trash floods into the Gulf of Honduras each year.

▶▶ Contributes to restoration and improvement of environment throughout Guatemala and to The Ocean Cleanup’s global mission of ridding the oceans of plastic.

Enjoy,

T. Scott Clendaniel

🥳 || #AI Training for 3.1 Million Worldwide || 109K Followers || Follow me for genuine ROI from Artificial Intelligence!

🔬 #Learning #AI:  𝗖𝗹𝗮𝘂𝗱𝗲 𝗖𝗼𝗱𝗲 𝗣𝗿𝗼𝗷𝗲𝗰𝘁 𝗦𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲!

PLEASE 𝗦𝗔𝗩𝗘 𝗮𝗻𝗱 𝗦𝗛𝗔𝗥𝗘 this so your team can take advantage of it!

Claude Code has absolutely EXPLODED and taken the lead by a country mile.

So please don't let poor project structure defeat you!  Fortunately, we have the amazing Brij kishore Pandey, the king of project structures, to show a better approach.


🧭

#AI #ArtificialIntelligece: Claude Code Project Structure!

🔬 AI Training: How to Build a Gemini 3 Pro Workflow App

Just watched Tianyu Xu create something remarkable. He built a fully functional node-based AI workflow app in just 10 minutes using Gemini 3 Pro. And here's the thing: this is exactly the kind of breakthrough moment we need to show our teams about what's possible with AI.

I want to give full credit to Tianyu for this incredible demonstration. His approach shows us something powerful about teaching staff to work with artificial intelligence. It's not about needing a computer science degree. It's about clear thinking and the right tools.

Here's how Tianyu built this workflow app. If you're training your team on AI tools, walk them through these steps:

𝗦𝘁𝗲𝗽 𝟭: 𝗚𝗲𝘁 𝘁𝗼 𝘁𝗵𝗲 𝗥𝗶𝗴𝗵𝘁 𝗣𝗹𝗮𝘁𝗳𝗼𝗿𝗺
Head to Google AI Studio and click "Build." This is where the magic starts. You're not writing code from scratch. You're describing what you want.

𝗦𝘁𝗲𝗽 𝟮: 𝗣𝗶𝗰𝗸 𝗬𝗼𝘂𝗿 𝗠𝗼𝗱𝗲𝗹
Select Gemini 3 Pro preview. This is the model doing the heavy lifting for you. It understands what you're asking for and can generate complete applications.

𝗦𝘁𝗲𝗽 𝟯: 𝗖𝗵𝗼𝗼𝘀𝗲 𝗬𝗼𝘂𝗿 𝗙𝗼𝘂𝗻𝗱𝗮𝘁𝗶𝗼𝗻
Select Nano Banana powered app. This gives you a starting point instead of a blank canvas, which makes the whole process faster.

𝗦𝘁𝗲𝗽 𝟰: 𝗪𝗿𝗶𝘁𝗲 𝗬𝗼𝘂𝗿 𝗕𝗹𝘂𝗲𝗽𝗿𝗶𝗻𝘁
Here's where your team's creativity matters. Write a prompt that spells out your key requirements. For Tianyu's workflow app, that meant a clean interface, nodes for text and image input, output nodes for text-to-image and image editing, resizable pieces, downloadable results, and keyboard shortcuts to speed things up.

𝗦𝘁𝗲𝗽 𝟱: 𝗕𝘂𝗶𝗹𝗱, 𝗧𝗲𝘀𝘁, 𝗗𝗲𝗯𝘂𝗴, 𝗮𝗻𝗱 𝗥𝗲𝗳𝗶𝗻𝗲
Watch what Gemini creates. It builds a working prototype. Then comes the real work. Testing it, finding what doesn't work quite right, and telling Gemini how to fix it. Tianyu discovered the app needed a crop tool to handle image aspect ratios better. He asked for that. Gemini added it. That back-and-forth is where learning happens.

This process took Tianyu about an hour of tweaking and testing until he had something that actually solved real problems. That's the honest part nobody talks about. But it's the part that matters for your team.

The takeaway? Your staff doesn't need to be programmers. They need to be clear about what they want and willing to test and adjust. That's the AI fluency you're really teaching them. Thanks again to Tianyu Xu for showing us this is truly possible.

- T. Scott Clendaniel

Sharing some PURE #Wisdom!  🥳






(Go follow Carmen Morin- she is constantly sharing gems 💎 like these!)

#1 LinkedIn Education Creator 🇨🇦 | Performance-Based Learning Strategist & Keynote Speaker | Concert Pianist Turned 7-Figure Education Founder

It's not about the money you offer. 

It's about the environment you create.

Most "retention strategies" completely miss the point.

Leaders obsess over compensation packages.

While their people dread Monday mornings, toxic meetings, and feeling invisible. 

A raise loses its shine in weeks.

But culture impacts you from the moment you wake up.

❌ Toxic cultures drain talent regardless of salary because they:

↳ Create stress that money can't compensate for 
↳ Limit growth potential and career advancement 
↳ Foster competition over collaboration 
↳ Ignore individual strengths and development needs 
↳ Prioritize output over human well-being

✅ Healthy cultures outperform pay raises because they:

↳ Build psychological safety where people can be authentic 
↳ Foster continuous learning and growth opportunities 
↳ Create meaningful connections between team members 
↳ Recognize contributions beyond just compensation 
↳ Align individual purpose with organizational mission 
↳ Support work-life integration, not just balance 
↳ Encourage experimentation without fear of failure

You can't buy your way out of toxic culture. 

You have to build your way out.

The best talent chooses environments, 
not just paychecks.

♻ Repost to champion culture-first leadership. 
🔔 Follow Carmen Morin for more on human-centered systems.

🔬 #AI: The Problem With All Those Plans to Replace Senior Executives

𝗡𝗲𝘄 𝗘𝗺𝗽𝗹𝗼𝘆𝗲𝗲 𝘃𝘀. 𝗮 𝗠𝗮𝘀𝘁𝗲𝗿 by Ravi Samrat Mishra 

So you have AI.  You're going to wipe out BOTH expensive existing talent AND new hires. You think you can just bring in agents to replace all that experience.

You might want to think again.  

Check out the video.  The younger employee is doing a heck of a lot of work.  In fact, he creates the crack that the "master" later exploits.

AI didn't swing a single hammer.  

YOU NEED to be onboarding the next generation!


𝗧𝗼𝗽 𝗧𝗶𝗽𝘀 𝗳𝗼𝗿 𝗧𝗿𝗮𝗶𝗻𝗶𝗻𝗴 𝗡𝗲𝘄 𝗘𝗺𝗽𝗹𝗼𝘆𝗲𝗲𝘀
-------------------------------------------

𝗦𝘁𝗮𝗿𝘁 𝗲𝗮𝗿𝗹𝘆: Begin training new hires on their first day so they are eager to learn and don't get lost to another organization. 
 
𝗖𝗿𝗲𝗮𝘁𝗲 𝗮 𝘁𝗿𝗮𝗶𝗻𝗶𝗻𝗴 𝗽𝗹𝗮𝗻: A plan with clear goals helps employees understand expectations and gives managers a consistent way to train new hires. 
 
𝗣𝗿𝗼𝘃𝗶𝗱𝗲 𝗿𝗲𝘀𝗼𝘂𝗿𝗰𝗲𝘀: Make training materials easily accessible and prepare onboarding resources. 
 
𝗨𝘀𝗲 𝘁𝗲𝗰𝗵𝗻𝗼𝗹𝗼𝗴𝘆: Create courses, test learning, and track progress using technology. 
 
𝗔𝘀𝘀𝗶𝗴𝗻 𝗮 𝗺𝗲𝗻𝘁𝗼𝗿 𝗼𝗿 𝗯𝘂𝗱𝗱𝘆: A mentor can help new employees feel valued and learn from someone already in the organization. 
 
𝗣𝗿𝗼𝘃𝗶𝗱𝗲 𝗿𝗲𝗴𝘂𝗹𝗮𝗿 𝗳𝗲𝗲𝗱𝗯𝗮𝗰𝗸: Feedback helps employees develop their skills and recognize their potential. 
 
𝗧𝗿𝗮𝗶𝗻 𝗼𝗻 𝗰𝗼𝗺𝗽𝗮𝗻𝘆 𝘃𝗮𝗹𝘂𝗲𝘀: Ensure new hires understand the company's vision, goals, and values. 
 
𝗘𝗻𝗰𝗼𝘂𝗿𝗮𝗴𝗲 𝘀𝗵𝗮𝗱𝗼𝘄𝗶𝗻𝗴: Allow new employees to shadow more experienced employees. 
 
𝗨𝘀𝗲 𝗼𝗻𝗯𝗼𝗮𝗿𝗱𝗶𝗻𝗴 𝗮𝘂𝘁𝗼𝗺𝗮𝘁𝗶𝗼𝗻: Onboarding automation can streamline and modernize the onboarding process. 

Credit: https://bit.ly/3O3cWn6

- T. Scott Clendaniel, AI Educator

LinkedIn Top Voice | Personal Branding Strategist & Executive Mentor | 81M+ Views | Helping Leaders & Founders Build Authority that Converts.

Hardest Part of Parenting?

Raising a human is the hardest work we will ever do. It is much more than just keeping them safe or making sure they have what they need. Our children are like little mirrors that reflect every single thing we do back at us. They do not just listen to our words, but they watch how we move through the world, even when we think they are not looking.

They see how we treat a stranger on the street and how we react when we are tired or frustrated. Our children mimic our morals and our behaviours every single moment of the day. They are learning from us even when we are not trying to teach them anything at all. It is a heavy weight to realise we are their primary map for life.

Every choice we make today becomes the voice inside their heads tomorrow. It makes me stop and look at my own actions with a lot of honesty. Am I being the kind of person I want them to grow up to be? We are not just raising kids, but we are shaping the future of how people treat one another.

How would you like them to be when they grow up?

#Parenting #Character #Legacy #Mindfulness #Inspiration

Children & Teen Personal Development Coach | I help students build self-belief/confidence through my Lead With Love program, empowering parents & teachers with leadership & emotional intelligence, grounded in Taekwondo.

View Ligia Chacón Hernández’s profile

Don’t Wait Too Long to Love ❤️

A little girl ran through the animal shelter, excited to find a dog.🐶

Kennel after kennel… barking, jumping, tails wagging.

Until she stopped.

In the corner sat an old brown dog with gray around his eyes. Quiet. Still.

His tag read: Milo — 10 years old.🐶

While the others barked for attention, Milo simply walked over and rested his head against the bars when she approached.

“I want him,” she said.

Her parents gently explained that Milo was very sick.
He might not have much time left.

She didn’t understand time.

She only understood that he had been waiting.

They promised to think about it and come back.

But the next morning, when they called the shelter…

Milo had passed away during the night.

The volunteer said he was found lying by the kennel door.

As if he was still waiting.

Sometimes the ones who wait the longest
are the ones we overlook the fastest.

Adopt the older dog.
Choose the one no one else sees.

Because love doesn’t need years.

Sometimes it just needs a chance. ❤️


Follow #Motivation #leadership #India #leadershiplove  #creativity #WhatInspiresme #education

#Advertisting #GenAI #Marketing 🎉👏🎉
⬇️⬇️⬇️  Most Stunning 3D Billboards Around the World!  ⬇️⬇️⬇️



𝗜𝗻𝗰𝗹𝘂𝗱𝗲𝘀:

 ▶▶ Audi
 ▶▶ BMW
 ▶▶ Coca-Cola
 ▶▶ Google 
 ▶▶ Louis Vuitton 
 ▶▶ Mercedes
 ▶▶ Netflix
 ▶▶ Nike
 ▶▶ NINTENDO WORLD™
 ▶▶ Spider-Man
 ▶▶ Super Mario Bros 
 ▶▶ VERSACE
 ▶▶ WhatsAPP


T. Scott Clendaniel

🔬#Learning #AI: Ten Commandments of AI Agents

How I wish I thought of these myself!  But I'm happy to share them.


The Commandments

➡️  Don't use agents just to "show off"
➡️  Agents need 2 characteristics: small, specialized, and decoupled
➡️  Enforce structured output
➡️  Explain the "why"
➡️  Orchestration is more important than autonomy 

➡️  Prompt Engineering Beats Fine-Tuning 
➡️  Tool 🔧 descriptions are VITAL
➡️  Cache like you mean it
➡️  Leverage shared artifacts
➡️  Log everything (seriously!)


Go support Jothi Moorthy!

-T. Scott Clendaniel

𝟏𝟎 𝐏𝐫𝐢𝐧𝐜𝐢𝐩𝐥𝐞𝐬 𝐄𝐯𝐞𝐫𝐲 𝐀𝐈 𝐄𝐧𝐠𝐢𝐧𝐞𝐞𝐫 𝐒𝐡𝐨𝐮𝐥𝐝 𝐅𝐨𝐥𝐥𝐨𝐰 𝐖𝐡𝐞𝐧 𝐁𝐮𝐢𝐥𝐝𝐢𝐧𝐠 𝐀𝐈 𝐀𝐠𝐞𝐧𝐭𝐬
Many teams rush into using AI agents without a clear strategy, often ending up with bloated, unreliable systems. Building powerful, production-grade AI agents requires discipline, structure, and thoughtful design choices. Here are 10 essential principles that separate hobby projects from scalable solutions:

𝟏. 𝐃𝐨𝐧’𝐭 𝐔𝐬𝐞 𝐀𝐠𝐞𝐧𝐭𝐬 𝐉𝐮𝐬𝐭 𝐭𝐨 𝐒𝐚𝐲 𝐘𝐨𝐮 𝐃𝐢𝐝
Use agents only where they truly add value. If a simple tool integration works, don’t overcomplicate it with orchestration.

𝟐. 𝐒𝐦𝐚𝐥𝐥, 𝐒𝐩𝐞𝐜𝐢𝐚𝐥𝐢𝐳𝐞𝐝, 𝐚𝐧𝐝 𝐃𝐞𝐜𝐨𝐮𝐩𝐥𝐞𝐝
Break big workflows into smaller specialized agents: Support, Manager, Verifier. Each should have a clear role.

𝟑. 𝐄𝐧𝐟𝐨𝐫𝐜𝐞 𝐒𝐭𝐫𝐮𝐜𝐭𝐮𝐫𝐞𝐝 𝐎𝐮𝐭𝐩𝐮𝐭
Always define output formats. This reduces ambiguity, improves reliability, and makes downstream processing easier.

𝟒. 𝐄𝐱𝐩𝐥𝐚𝐢𝐧 𝐭𝐡𝐞 𝐖𝐡𝐲, 𝐍𝐨𝐭 𝐉𝐮𝐬𝐭 𝐭𝐡𝐞 𝐖𝐡𝐚𝐭
Agents should understand context. When they know the why, they make better decisions.

𝟓. 𝐎𝐫𝐜𝐡𝐞𝐬𝐭𝐫𝐚𝐭𝐢𝐨𝐧 𝐎𝐯𝐞𝐫 𝐁𝐥𝐢𝐧𝐝 𝐀𝐮𝐭𝐨𝐧𝐨𝐦𝐲
Orchestrate multiple steps intelligently instead of relying on a single agent to do everything. This boosts flexibility and reduces failure.

𝟔. 𝐏𝐫𝐨𝐦𝐩𝐭 𝐄𝐧𝐠𝐢𝐧𝐞𝐞𝐫𝐢𝐧𝐠 > 𝐅𝐢𝐧𝐞-𝐓𝐮𝐧𝐢𝐧𝐠
Maximize what you can do with retrieval, RAG, prompt patterns, and few-shot techniques before reaching for expensive fine-tuning.

𝟕. 𝐃𝐨𝐮𝐛𝐥𝐞 𝐃𝐨𝐰𝐧 𝐨𝐧 𝐓𝐨𝐨𝐥 𝐃𝐞𝐬𝐜𝐫𝐢𝐩𝐭𝐢𝐨𝐧𝐬
Agents need crystal clear instructions on how and when to use tools, including what to avoid and constraints to follow.

𝟖. 𝐂𝐚𝐜𝐡𝐞 𝐋𝐢𝐤𝐞 𝐘𝐨𝐮 𝐌𝐞𝐚𝐧 𝐈𝐭
Avoid unnecessary calls. Cache results and orchestrate smartly to reduce costs and latency.

𝟗. 𝐔𝐬𝐞 𝐒𝐡𝐚𝐫𝐞𝐝 𝐀𝐫𝐭𝐢𝐟𝐚𝐜𝐭𝐬
Use shared document IDs or structured outputs to avoid data duplication and improve collaboration.

𝟏𝟎. 𝐋𝐨𝐠 𝐄𝐯𝐞𝐫𝐲𝐭𝐡𝐢𝐧𝐠 𝐒𝐞𝐫𝐢𝐨𝐮𝐬𝐥𝐲
Good logging makes debugging, optimization, and scaling infinitely easier. If it’s not logged, it did not happen.

𝐓𝐡𝐞 𝐝𝐢𝐟𝐟𝐞𝐫𝐞𝐧𝐜𝐞 𝐛𝐞𝐭𝐰𝐞𝐞𝐧 𝐚 𝐛𝐚𝐬𝐢𝐜 𝐚𝐠𝐞𝐧𝐭 𝐚𝐧𝐝 𝐚 𝐬𝐜𝐚𝐥𝐚𝐛𝐥𝐞 𝐬𝐲𝐬𝐭𝐞𝐦 𝐥𝐢𝐞𝐬 𝐢𝐧 𝐭𝐡𝐞𝐬𝐞 𝐝𝐞𝐬𝐢𝐠𝐧 𝐜𝐡𝐨𝐢𝐜𝐞𝐬. 𝐖𝐡𝐢𝐜𝐡 𝐨𝐟 𝐭𝐡𝐞𝐬𝐞 𝐩𝐫𝐢𝐧𝐜𝐢𝐩𝐥𝐞𝐬 𝐝𝐨 𝐲𝐨𝐮 𝐚𝐥𝐫𝐞𝐚𝐝𝐲 𝐟𝐨𝐥𝐥𝐨𝐰 𝐢𝐧 𝐲𝐨𝐮𝐫 𝐛𝐮𝐢𝐥𝐝𝐬?

♻️ Repost this to help your network get started
➕ Follow Jothi for more

Anthony Alcaraz

Writes on ML/LLM Ops 🎨 | Follow me (4,8 K+) 🔔

I help AI startups go from promising to unstoppable.

I also write about it, my book Agentic Graph RAG (O'Reilly, 2026) is the practical blueprint for building AI systems that combine retrieval, reasoning, and autonomous decision-making at enterprise scale.

https://www.oreilly.com/library/view/agentic-graph-rag/9798341623163/

My career has been a journey across both sides of the table: building AI products as a CTO, transforming legacy systems as a Field CTO for enterprises and government, and now at AWS, helping the most exciting GenAI startups scale across Europe and the US.

I've shipped GenAI platforms, deployed MLOps pipelines serving millions, advised L'Oréal on AI-for-BI transformation, and lectured at Oxford on the intersection of LLMs and graphs. But what gets me out of bed is working with founders who are solving hard problems with bold ideas.

At AWS, I connect high-potential startups with the resources, partnerships, and enterprise pathways they need to grow. Before that, I was in the trenches, building, breaking, and shipping.

If you're a founder, VC, or enterprise leader thinking about GenAI, agentic systems, or knowledge graphs, I'd love to chat.

Chief AI Officer & Decision scientist

Management in AI Projects and Engineering Teams:
In my managerial role, I oversee AI projects and lead teams of engineers in crafting cutting-edge solutions. My focus is on driving strategic initiatives, ensuring projects are aligned with organizational goals, and fostering a culture of innovation and excellence. I take a hands-on approach to problem-solving and mentorship, ensuring that both projects and people reach their fullest potential.

Development in LLM/ML Ops:
My development expertise lies in the specialized domains of Large Language Models Operations (LLM Ops) and Machine Learning Operations (ML Ops). I am responsible for streamlining machine learning workflows from development to deployment. My work enables organizations to not only adopt but excel in leveraging machine learning technologies, facilitating continuous learning and adaptability across the board.

Marketing and Strategic Partnerships:
In the realm of marketing, I focus on building and nurturing strategic partnerships. My efforts are geared towards creating synergies with like-minded organizations and industry leaders. By forming valuable partnerships, I amplify brand visibility, foster innovation, and open doors to new business opportunities. The result is a win-win situation for all stakeholders involved, enhancing both market reach and impact.

Three engineers at OpenAI shipped 1 million lines of code without writing a single line themselves. 🔺

Separately, a 7B parameter model beat GPT-4o by 41.9% on multi-step agent tasks.

Both teams discovered the same principle: the environment determines the agent's ceiling, not the model.

@Ryan Lopopolo's Harness team ran Codex agents for 5 months. 1,500 PRs merged. 3.5 PRs per engineer per day. The constraint was deliberate: zero human code. When an agent failed, the fix was never "try harder." The question was always: what capability is missing, and how do we make it legible?

They built progressive disclosure architectures where AGENTS.md stays under 100 lines and points to structured docs. Per-worktree isolation with full observability stacks torn down after each task. Background garbage collection agents scanning for pattern drift on a regular cadence. Agents reviewing other agents' PRs, humans optional.

SkillRL (Xia et al., 2026) attacked the same problem from the opposite direction. Instead of engineering the environment upfront, they let agents distill their own experiences into reusable skills. Raw trajectory storage caused a 25.2% performance drop. Agents that memorized their past got worse. Agents that abstracted their past into structured skills (name, principle, whentoapply) improved continuously. The library grew from 55 to 100 skills during training through recursive co-evolution with GRPO reinforcement learning.

Ashpreet B.  Bedi's open-source Dash framework proved the same pattern works without fine-tuning at all: context plus memory beats retraining. OpenAI's own Kepler data agent runs on the same principle across 3,500 users and 600 petabytes.

Both approaches converge on the same two bets: structured knowledge beats raw data, and mechanical enforcement beats advisory guidelines. Their systems co-evolve rather than ship-and-forget.

The gap neither has solved: governance. SkillRL uses flat semantic retrieval with no conflict detection between skills. Harness encodes golden principles manually. Neither tracks provenance, dependency ordering, or compositional constraints. Skills form dependency graphs with conflict edges, composition paths, and provenance chains. The governance layer for self-improving agents is a graph structure.

The engineer's job has shifted. Write zero code. Build the environment. Let the graph enforce the rules.

And not only engineers…

Scaling Agentic Startups to Enterprise @AWS | Author of Agentic Graph RAG (O’Reilly) | Business Angel | Supreme Commander of Countless Agents

No alternative text description for this image

taplio

The Battle

T. Scott Clendaniel

Anthony Alcaraz

Followers

Posts

Engagements

Likes

Performance ScoreT. Scott Clendaniel's LinkedIn Performance

Performance ScoreAnthony Alcaraz's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

T. Scott Clendaniel

Anthony Alcaraz

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);}T. Scott Clendaniel's LinkedIn Performance

Performance ScoreAnthony Alcaraz's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreT. Scott Clendaniel's LinkedIn Performance