THE NEXT BIG THING: AI & WEB3 (The Future of Tech Careers).

TechCrushHQ avatar
TechCrush.pro
@TechCrushHQSep 19, 2025

The Spaces featured a practical conversation on starting and advancing careers in AI and Web3. Hamed Habasha (AI/Robotics, South Korea) traced his path from mechanical engineering and backend programming to embedded systems and machine learning during his master’s in Egypt, emphasizing mentorship, prompt engineering, and today’s accessible no/low-code and cloud AI tools. Steven Obmira (AI/Robotics, Head of Innovations, Nigeria) shared how structured learning and problem-driven projects sustained motivation, with concrete examples applying AI across retail (data chatbots), robotics (computer vision with YOLO/Cascade), and agriculture (nutrient optimization). Ola Hamid (smart contract developer) demystified Web3 beyond crypto—highlighting zero-knowledge proofs and diverse roles (security, data, legal, design)—and stressed curiosity, repetition, structure, and community. The speakers agreed AI won’t replace workers who adapt; integrating AI accelerates delivery. For newcomers, two routes emerged: use existing AI tools effectively (prompting) or become a builder (learn Python, algorithms, cloud). Community and mentorship reduce overwhelm and create momentum. Looking ahead, they forecast tighter human–AI partnerships via agents and advised going deeper technically to stay relevant. A robust Q&A covered prerequisites, building an LLM-powered chatbot, differences between data science, ML, and AI, gear needs, and program logistics.

Future of Tech Careers — AI and Web3 (Tech Fresh Mentorship Sessions)

Speakers and roles

  • Host: Tech Fresh mentorship sessions facilitator (name not clearly stated in the recording)
  • Hamed (AI/Robotics, based in South Korea): AI and robotics practitioner with a mechanical engineering background; experience in embedded systems and applied AI research across domains
  • Ola Hamid (Smart contract/Web3 developer): Smart contract and blockchain developer; works on security; teaches Web3; co-founder/smart contract developer for a centralized exchange; mentors newcomers
  • Steven Obmira (AI/Robotics & Innovation): Robotics engineer and AI practitioner; Head of Innovations at RAIN Nigeria; alumnus of the Microsoft Startup program (supported by Microsoft and OpenAI); applies AI across retail, healthcare, agriculture, and robotics

Session focus

  • Demystifying AI and Web3 for newcomers and career transitioners
  • Practical learning paths, tools, and strategies to maintain motivation
  • The future of work in AI and Web3 and how to stay relevant

Hamed’s journey and guidance (AI & Robotics)

  • Origin story
    • Started as a backend developer/programmer during/after undergraduate studies; languages used included PHP, HTML; simultaneously explored electronics and embedded systems (microcontrollers, serial comms, soldering) while studying mechanical engineering
    • Mentorship moments were pivotal: a lecturer returning from the Czech Republic nudged him toward C/C++ in embedded contexts; later, casual insights from a peer “connected the dots” after a year of heavy but fragmented self-study
    • Adopted a project mindset early: integrated mechanics, electronics, and programming in his undergraduate project to set his future trajectory
  • Transition into AI
    • Discovered machine learning/AI during his master’s (in Egypt) through a Japanese professor’s project-based learning class and a roommate researching human activity recognition (e.g., smartwatch movement classification)
    • Struggled initially with the mathematical rigor (research-level derivations) and mapping theory to code; resolved via targeted mentorship and a course that connected math, code, and implementation
    • Principle: You can use AI without deep math, but mathematical depth determines how far you can push innovation and make original contributions
  • Tooling and accessibility today
    • Emphasized that entry is much easier now: cloud providers (Google Cloud, AWS, others) offer simplified ML stacks; examples include using SQL-driven pipelines with integrated model training (e.g., BigQuery ML–style experiences)
    • Even non-traditional pathways (JavaScript, SQL) can now contribute to AI workflows
    • Prompt engineering is a key skill for non-coders and coders alike; learning personas and structured prompting substantially improves LLM outcomes
  • Advice for beginners
    • Don’t be discouraged by hype or jargon; focus on getting correct information and building intuition via guided learning
    • Seek mentors to accelerate progress; you can learn alone, but it’s slower and more error-prone in fast-moving fields
    • Consider no-code/low-code platforms for AI as a bridge if programming is a barrier, while acknowledging that programming unlocks greater depth and autonomy

Steven’s journey and guidance (AI/Robotics & Applied AI)

  • Origin story
    • Early tinkerer in secondary school; strong programming start aided by a good computer lab; also pursued design to fund learning
    • In university (Engineering Physics), he pursued robotics with structured, institutional learning to avoid fragmented self-study
  • First practical AI moment
    • In robotics, replaced complex inverse kinematics math with a trained ML model to predict end-effector (x, y, z) from joint values—an eye-opener about AI as function approximation
  • Applying AI beyond robotics
    • Expanded into deep learning/computer vision and applied AI for business value
    • Guiding philosophy: find a real problem you’re passionate about solving; let that end goal anchor your learning trajectory
  • Motivation and challenges
    • Challenges: Cost of gear/internet/power/tuition; motivation dips from overwhelm; solved by working multiple gigs (design) to fund learning and by anchoring progress to a real problem/goal
    • Structured curricula help avoid scattered learning and accelerate relevance

Practical “day in the life” and real project workflows (Steven)

  • Retail/SME analytics assistant
    • Goal: help SMEs access/understand their data (sales, customers, costs, suppliers) via natural-language interfaces
    • Workflow: define goals → data gathering (sales, churn, costs) → cleaning and feature engineering → model selection/training/evaluation (confidence scoring) → deployment (e.g., web app or integration with WhatsApp) → UX that feels like “chatting with your business”
  • Robotics (object detection for manipulation)
    • Goal: robot arm recognizes objects on a table and arranges them
    • Workflow: collect images → annotate/label (bounding boxes; classes: pencil, pen, etc.) → choose model families (e.g., Haar Cascades, YOLO) → train/validate → deploy to edge (e.g., Raspberry Pi/mini-computer) → integrate detections with motion planning to actuate the arm
  • Agriculture (tomato growth optimization)
    • Goal: accelerate growth via correct nutrient composition
    • Approach: computer vision to diagnose leaf nutrient deficiencies → recommend tailored nutrient mixes via ML models → feedback loop into growing system

Ola Hamid’s journey and guidance (Web3)

  • Origin story
    • Mechanical engineering student initially “lost”; curiosity about blockchain sparked by a crypto link experience (even losing money early didn’t deter him)
    • Started in community/support roles, then moved into technical learning—took courses, built small projects, iterated
    • Now a smart contract/blockchain developer focused on security; co-founder/smart contract developer for a centralized exchange; teaches Web3 development
  • Accessibility and mindset
    • Repetition and curiosity are differentiators; anyone can enter Web3 with the right mindset and consistent practice
    • Structured curricula accelerate progress; communities/mentors shorten the time to resolve blockers (even simple setup issues can otherwise stall progress)
  • Web3 is broader than crypto
    • Highlighted zero-knowledge proofs (ZK) as a non-crypto example—verifiable claims without revealing underlying data (e.g., more trustworthy elections auditing)
    • Emphasized the need for education to connect Web3 tech to real-world problems; non-obviousness to the public is a key adoption challenge
  • Challenges and how to overcome
    • Web3 often feels abstract relative to daily life; solution is better storytelling, real use cases, and targeted education
    • Community and mentorship are essential—don’t learn in isolation
  • Career positioning and next 2–5 years
    • Be courageous and define a clear goal; pick a niche aligned to your interests: security auditing, data engineering, legal (Web3-savvy law), design, technical writing, etc.
    • Map your existing domain skills to Web3 to discover differentiated opportunities

Is AI going to take jobs?

  • Hamed’s perspective
    • Analogy: news helicopters replaced by drones—pilots weren’t eliminated; roles evolved
    • AI is a permanent shift. It may not “take your job” outright, but failing to integrate AI into your workflow will phase you out as peers leverage AI to deliver faster and better
    • Productivity leap: tasks that took weeks can compress into days if you know how to prompt and integrate AI effectively

Staying relevant in 3–5 years (AI and Web3)

  • Steven’s view on AI
    • Expect tighter human–AI partnerships via autonomous/agentic systems that perform actions across your tools and workflows
    • The differentiator: be technical. Beyond prompt engineering, understand the mathematical underpinnings, benchmarking metrics, model behavior, and systems integration to outperform superficial users
  • General career advice (across AI/Web3)
    • Depth wins. Many can use basic tools; few can architect systems, reason about trade-offs, and improve models or protocols
    • Anchor your learning to a real problem in a domain you care about; it sustains motivation

Community, mentorship, and structured learning

  • All speakers endorsed mentorship and community
    • Faster progress, fewer dead-ends; peers can unblock you on seemingly trivial issues that otherwise consume days
    • Structured curricula counter scattered information and provide a coherent path
  • The host highlighted Tech Fresh’s Boot Camp approach
    • Beginner-friendly curriculum; multiple mini-projects culminating in a real-world capstone
    • Emphasis on problem selection early—decide the real-world problem you’ll solve to maintain momentum
    • Program details and curriculum to be shared with Scholarship Boot Camp applicants; track switching allowed based on informed interest

Q&A highlights

  • How to start AI without a technical background?
    • Steven: Two paths
      • Tool user: start now with LLMs (ChatGPT/Claude), practice prompt engineering, try cloud tools; no heavy math required
      • System builder: learn programming (Python recommended), then algorithms/ML; depth comes with math and hands-on projects
    • Hamed: No-code AI platforms can bridge the gap; programming remains the best route for autonomy, but structured prompting and no-code tools can deliver meaningful results and momentum
  • Building a domain chatbot (dog breeding advisor)
    • Steven: Architecture blueprint
      • Front end: chat UI (HTML/CSS/JS, React, or Flutter)
      • Backend: Python (Flask/FastAPI) or Node.js (Express); receive user query
      • Model: call OpenAI/Claude via API key; return model response to UI
      • Iterate on prompts, add retrieval augmentation if needed; evaluate answers for reliability
  • Can AI help non-technical research fields (history/international studies)?
    • Hamed: Yes. LLMs trained on large corpora (e.g., Wikipedia, research indices). Use research-friendly tools (e.g., Perplexity) and learn prompt engineering to improve sourcing and traceability. Boot Camp will cover fundamentals
  • Data Science vs. Machine Learning vs. AI (Steven)
    • Data Science: gathering, cleaning, visualizing, and exploring data
    • Machine Learning: learning mathematical relationships/patterns to predict or classify
    • AI: productized systems that combine ML/DL (and sometimes rule-based systems) to act autonomously or assistively
  • Web3 equipment and entry prerequisites
    • Ola Hamid: To learn concepts, a basic internet-capable device is fine. To develop and deploy, you’ll need a laptop that can run a code editor (e.g., VS Code) and developer tooling. Any reasonable modern machine is acceptable
  • Mentorship asks
    • Ola Hamid is open to mentoring and highlighted the webinar’s role in onboarding; Boot Camp offers deeper mentorship structures

Actionable takeaways

  • For absolute beginners (AI)
    • Start with LLMs and prompt engineering; pick a domain problem to drive learning
    • Try cloud ML tooling with simplified interfaces; explore no-code options to build momentum
    • If aiming for technical depth, begin Python now and implement small ML projects end-to-end
  • For absolute beginners (Web3)
    • Learn blockchain basics, then pick a niche (security, data engineering, legal, design)
    • Follow a structured curriculum; build small dApps/contracts; join a community for code reviews and support
  • For everyone
    • Choose a real problem (business/healthcare/agri/logistics/civic tech) and work backward to the skills needed
    • Seek mentors and communities; don’t isolate
    • Invest in fundamentals (math/programming for AI; protocol/security basics for Web3) to stand out
    • Expect agentic AI and broader Web3 applications (e.g., ZK) to reshape workflows—prepare to integrate these into your toolset

Closing notes

  • The host emphasized:
    • Curiosity and grit are essential; you don’t need to be an engineer to start, but depth and consistency keep you competitive
    • The Boot Camp curriculum (beginner-friendly with progressive projects) will be shared to applicants; laptop guidance and track changes are supported via the community
    • Upcoming sessions will cover additional tracks (e.g., mobile/game development)