Becoming An AI Practitioner
Career Paths in AI
A common misconception among newcomers: "working in AI" means exclusively training neural networks. In reality, the field spans a wide range of roles, many of w
JrCodex·6 min read
Jr Codex AI Notes
Level: Intermediate Prerequisites: Module 6: AI Ethics, Safety & Society Time to complete: ~20 minutes
Table of Contents
- The AI Job Market Is Broader Than "Machine Learning Engineer"
- Common AI-Related Roles
- Where Classical AI Knowledge (This Curriculum) Actually Shows Up
- Skills That Transfer Across Every Role
- Building a Portfolio
- A Realistic Learning Path Forward
- Summary & Next Steps
1. The AI Job Market Is Broader Than "Machine Learning Engineer"
A common misconception among newcomers: "working in AI" means exclusively training neural networks. In reality, the field spans a wide range of roles, many of which lean heavily on the classical, symbolic AI techniques covered throughout this curriculum — not just the learning-based methods in the ML/DL/NLP notes.
2. Common AI-Related Roles
A Non-Exhaustive Map of AI-Related Roles
─────────────────────────────────────────
Machine Learning Engineer: builds and deploys trained
models — heavy overlap with
the Machine Learning and
Deep Learning Notes
Research Scientist: develops new algorithms/
techniques — draws on
this ENTIRE curriculum's
theoretical foundations
AI/Data Scientist: applies AI/ML techniques
to specific business
questions — heavy
overlap with the Data
Science Notes
Robotics/Automation Engineer: relies heavily on
THIS curriculum's
Module 2 (search),
Module 3 (planning),
and Module 5 (RL)
AI Ethics/Policy Specialist: draws directly
on Module 6's
content —
bias, fairness,
governance
Prompt Engineer / AI Application draws on the
Developer: NLP/LLM Notes,
building
products ON
TOP OF existing
LLMs rather
than training
models from
scratch
MLOps/AI Infrastructure Engineer: focuses on
reliably
deploying
and
monitoring
AI systems
at scale
─────────────────────────────────────────
Notice how few of these roles require deep neural network expertise as their primary skill — many lean heavily on exactly the classical techniques (search, planning, probabilistic reasoning, decision theory) built across this curriculum's Modules 2-5.
3. Where Classical AI Knowledge (This Curriculum) Actually Shows Up
Real-World Applications of THIS Curriculum's Specific Content
─────────────────────────────────────────
Module 2 (Search): route planning (logistics, delivery
apps), puzzle/game AI, resource
scheduling, network routing
Module 3 (Knowledge/ expert systems in regulated
Planning): industries (Module 3, Ch.6),
robotics task planning,
business rule engines
(still widely used, per
Module 3, Ch.6's "where rule-
based AI still matters")
Module 4 (Uncertainty): fraud detection, medical
diagnosis support, spam
filtering, any system
requiring principled
confidence estimates
Module 5 (Agents/RL): robotics control,
resource allocation,
recommendation systems,
game AI, autonomous
vehicle decision-making
Module 6 (Ethics/Safety): increasingly a
DEDICATED role
(AI ethics/policy/
trust & safety) at
any company
deploying AI at scale
─────────────────────────────────────────
A genuinely valuable, differentiating skill: many AI practitioners today are strong in ML/DL/NLP but have never formally studied search, planning, or classical probabilistic reasoning — having this curriculum's foundation is a real, practical advantage for roles involving optimization, scheduling, robotics, or any problem that doesn't reduce cleanly to "train a model on labeled data."
4. Skills That Transfer Across Every Role
Durable Skills, Regardless of Specific Role
─────────────────────────────────────────
Problem Formulation: the ability to frame a messy
real-world problem as a search
problem, CSP, MDP, or another
formal structure (Module 2-5's
entire approach) is MORE
valuable, long-term, than
knowing any single algorithm
Evaluating Trade-offs: every technique in this
curriculum involved genuine
trade-offs (BFS vs DFS's
memory/optimality, symbolic
vs learned models'
transparency/flexibility) —
recognizing and articulating
these trade-offs is a skill
that transfers to ANY new
technique you encounter later
Responsible Deployment Judgment: Module 6's content — knowing
WHEN to demand explainability,
check for bias, or involve
legal/ethics review — is
increasingly a baseline
expectation, not a
specialist skill
Communicating Technical Work: directly connects to the
Data Science Notes'
storytelling chapter —
every role above requires
explaining technical
decisions to non-technical
stakeholders
─────────────────────────────────────────
5. Building a Portfolio
Concrete Ways to Demonstrate This Curriculum's Skills
─────────────────────────────────────────
- Implement and compare BFS/DFS/A* on a real routing problem
with actual map data (Module 2)
- Build a Sudoku or N-Queens solver using CSP techniques
with backtracking + constraint propagation (Module 2, Ch.6)
- Build a small expert system for a domain you know well
(Module 3, Ch.6) — even a hobby domain works well as a
portfolio piece
- Implement Q-learning (Module 5, Ch.5) on a simple custom
game/gridworld environment
- Write up a bias/fairness AUDIT of a public dataset or
model (Module 6, Ch.1) — a genuinely valuable, differentiated
portfolio piece few candidates think to include
─────────────────────────────────────────
Chapter 2's capstone project is designed to be exactly this kind of portfolio-ready piece — a complete, working program combining several of this curriculum's core techniques.
6. A Realistic Learning Path Forward
Where to Go From Here, Depending on Career Direction
─────────────────────────────────────────
Interested in ML Engineering/Data Science roles:
→ Machine Learning Notes, then Deep Learning Notes
Interested in NLP/LLM-focused roles:
→ Deep Learning Notes (neural network foundations first),
then NLP/LLM Notes
Interested in Robotics/Automation:
→ Deepen Module 2 (search) and Module 5 (RL) further with
dedicated robotics/control theory resources beyond this
curriculum's introductory scope
Interested in AI Ethics/Policy roles:
→ Module 6 is your foundation; supplement with dedicated
policy/law resources specific to your jurisdiction of
interest (Module 6, Ch.5's honest limitation applies
directly here)
Undecided, want broad competence:
→ Complete Machine Learning, then Deep Learning, then
NLP/LLM Notes in sequence — the standard, well-rounded
path this entire curriculum is built around
─────────────────────────────────────────
7. Summary & Next Steps
Key Takeaways
- The AI job market spans far more roles than "train neural networks" — many roles (robotics, business rule systems, fraud detection, AI ethics/policy) lean heavily on this curriculum's classical techniques specifically.
- Problem formulation, trade-off evaluation, responsible deployment judgment, and clear communication are durable skills that transfer across every AI-related role, regardless of the specific techniques in fashion.
- A strong portfolio demonstrates hands-on implementation of this curriculum's techniques (search, CSPs, expert systems, Q-learning, bias audits) — concrete, working artifacts, not just conceptual familiarity.
- The right next step in this curriculum depends on career direction — but Machine Learning → Deep Learning → NLP/LLM is the standard, well-rounded path for most learners.
Concept Check
- Name two AI-related roles that rely heavily on classical (non-learning-based) AI techniques from this curriculum.
- Why is "problem formulation" described as more durably valuable than knowing any single specific algorithm?
- What's a concrete, portfolio-worthy project you could build using Module 2's search techniques?
Next Chapter
→ Chapter 2: Capstone Project — A Puzzle-Solving Agent
Jr Codex — 1-on-1 Personalized Coaching | Back to Module Index