Role-Specific Prep Paths
Reading time: ~20 min | Interview relevance: Critical | Roles: All AI/ML roles
The Real Interview Moment
You have just decided to make the leap. Maybe you saw a job posting that excited you, got a recruiter message on LinkedIn, or simply decided it is time for a career change into AI and ML. You open your browser, search "how to prepare for ML interviews," and immediately drown in a sea of advice: study LeetCode, read papers, learn system design, practice behavioral questions, build projects, contribute to open source...
The overwhelm is real. And the most dangerous response to overwhelm is doing a little bit of everything and mastering nothing.
Here is the truth that seasoned AI professionals know: the skills that matter most depend entirely on the specific role you are targeting. A Machine Learning Engineer interview looks nothing like a Data Scientist interview, which looks nothing like an MLOps interview. The candidate who wins is not the one who studied the most -- it is the one who studied the right things for their target role.
This section of the handbook is your GPS. Instead of wandering through a jungle of preparation materials, you will follow a structured, week-by-week plan designed specifically for the role you want. Each path has been refined based on hundreds of real interview experiences and hiring manager feedback.
Why Role-Specific Preparation Matters
:::tip The 80/20 Rule of Interview Prep 80% of your interview success comes from 20% of the preparation material -- but which 20% depends entirely on the role. A Data Scientist who spends weeks on Kubernetes is wasting time. An MLOps Engineer who ignores infrastructure design is making a critical mistake. :::
The AI/ML Role Landscape
The AI/ML industry has matured significantly. What used to be a single "data scientist" role has fragmented into highly specialized positions, each with distinct interview processes:
How Interview Processes Differ by Role
| Dimension | MLE | AI Engineer | MLOps | Data Scientist | Research Engineer | Data Engineer |
|---|---|---|---|---|---|---|
| Coding weight | 30% | 25% | 25% | 20% | 25% | 30% |
| ML theory | 25% | 15% | 20% | 25% | 35% | 5% |
| System design | 25% | 20% | 35% | 10% | 10% | 30% |
| Behavioral | 20% | 15% | 20% | 20% | 10% | 15% |
| Domain-specific | -- | 25% (LLMs) | -- | 25% (stats) | 20% (math) | 20% (SQL) |
| Typical prep time | 8 weeks | 6 weeks | 8 weeks | 6 weeks | 10 weeks | 6 weeks |
| Hardest part | ML system design | LLM depth | Infra breadth | Business sense | Math rigor | Scale thinking |
How to Choose Your Path
Choosing the right prep path is itself one of the most important decisions you will make. Here is a structured decision framework.
Step 1: Assess Your Background
Step 2: Match Your Strengths to Roles
Choose MLE if you:
- Enjoy both coding and ML theory equally
- Want to build models that serve millions of users
- Are comfortable with ambiguous, open-ended problems
- Have experience with Python, ML frameworks, and production systems
Choose AI Engineer if you:
- Are excited about LLMs, agents, and generative AI
- Prefer building applications over training models
- Enjoy API design, prompt engineering, and integration work
- Want to work at the cutting edge of applied AI
Choose MLOps if you:
- Love infrastructure, DevOps, and systems thinking
- Enjoy making other engineers more productive
- Are comfortable with Docker, Kubernetes, and cloud platforms
- Care more about reliability and scalability than model architecture
Choose Data Scientist if you:
- Have strong statistics and probability foundations
- Enjoy translating business questions into data analyses
- Are comfortable presenting findings to non-technical stakeholders
- Love A/B testing, experimentation, and causal inference
Choose Research Engineer if you:
- Are passionate about understanding how and why algorithms work
- Enjoy reading and implementing academic papers
- Have strong mathematical foundations (linear algebra, optimization, probability)
- Want to push the boundaries of what is possible in AI
Choose Data Engineer if you:
- Love designing efficient data systems and pipelines
- Are passionate about SQL, data modeling, and distributed computing
- Enjoy building the foundation that data scientists and ML engineers rely on
- Care about data quality, reliability, and governance
Step 3: Consider the Market
:::note Current Market Trends (2025-2026)
- AI Engineer roles are growing fastest, driven by LLM adoption
- MLE remains the most consistently in-demand role
- MLOps is increasingly valued as companies mature their ML practices
- Data Scientist roles are evolving -- many now require ML engineering skills
- Research Engineer roles are concentrated at top labs and large tech companies
- Data Engineer demand remains strong as data infrastructure grows more complex :::
The Six Prep Paths at a Glance
Path 1: Machine Learning Engineer (8 Weeks)
Who it is for: Software engineers transitioning to ML, or ML practitioners aiming for top-tier MLE roles.
Time commitment: 3-4 hours/day on weekdays, 5-6 hours on weekends
Core focus areas:
- Coding (30%): LeetCode medium/hard, ML-specific coding
- ML Fundamentals (25%): Classical ML, deep learning theory, evaluation metrics
- System Design (25%): End-to-end ML system design, feature stores, model serving
- Behavioral (20%): ML project stories, impact narratives, leadership principles
What makes MLE prep unique: You need to be equally strong in software engineering AND machine learning. This is the most balanced and arguably the most demanding prep path.
Start here: MLE Prep Path
Path 2: AI Engineer (6 Weeks)
Who it is for: Software engineers looking to specialize in LLM-powered applications, or ML engineers pivoting to the AI engineering space.
Time commitment: 3-4 hours/day on weekdays, 4-5 hours on weekends
Core focus areas:
- LLM Focus (40%): Transformer internals, prompting, fine-tuning, RAG, agents
- Coding (25%): Python, API design, async programming, LLM integration patterns
- System Design (20%): LLM serving infrastructure, RAG architectures, agent systems
- Behavioral (15%): AI product sense, ethical AI, project impact stories
What makes AI Engineer prep unique: This is the newest role in the landscape and interview formats are still evolving. You need deep LLM knowledge combined with strong software engineering skills.
Start here: AI Engineer Prep Path
Path 3: MLOps Engineer (8 Weeks)
Who it is for: DevOps engineers moving into ML, or ML engineers who want to focus on infrastructure and operationalization.
Time commitment: 3-4 hours/day on weekdays, 5-6 hours on weekends
Core focus areas:
- System Design (35%): ML pipelines, model serving, monitoring, CI/CD for ML
- Coding (25%): Python, infrastructure-as-code, scripting, automation
- ML Fundamentals (20%): Enough ML to understand what you are operationalizing
- Behavioral (20%): Reliability stories, incident response, cross-team collaboration
What makes MLOps prep unique: You need the broadest infrastructure knowledge of any role. Docker, Kubernetes, cloud services, monitoring, logging, CI/CD -- the surface area is enormous.
Start here: MLOps Prep Path
Path 4: Data Scientist (6 Weeks)
Who it is for: Analysts moving into data science, or academics transitioning to industry data science roles.
Time commitment: 2-3 hours/day on weekdays, 4-5 hours on weekends
Core focus areas:
- Statistics and Probability (25%): Hypothesis testing, distributions, Bayesian thinking
- SQL and Data Manipulation (25%): Complex queries, window functions, pandas
- ML Fundamentals (20%): Classical ML, feature engineering, model evaluation
- Business Case Studies (15%): Product sense, metric design, A/B testing
- Behavioral (15%): Communication, stakeholder management, impact stories
What makes Data Scientist prep unique: You must excel at communication. The best data scientists translate complex statistical findings into actionable business insights. Technical chops alone will not get you the offer.
Start here: Data Scientist Prep Path
Path 5: Research Engineer (10 Weeks)
Who it is for: Engineers with strong math backgrounds aiming for research labs, or PhD students preparing for industry research positions.
Time commitment: 4-5 hours/day on weekdays, 6-8 hours on weekends
Core focus areas:
- Paper Reading and Implementation (35%): Read, understand, and implement seminal papers
- Mathematical Foundations (20%): Linear algebra, optimization, probability theory, information theory
- Coding (25%): Algorithm implementation, efficient numerical computing, PyTorch internals
- Research Taste and Presentation (10%): Identifying promising research directions, presenting findings
- Behavioral (10%): Research collaboration, handling failure, scientific rigor
What makes Research Engineer prep unique: This is the longest and most intellectually demanding path. You are not just learning to use ML -- you are learning to advance it. Expect to read and implement 2-3 papers per week.
Start here: Research Engineer Prep Path
Path 6: Data Engineer (6 Weeks)
Who it is for: Backend engineers interested in data infrastructure, or analysts wanting to move into engineering-focused roles.
Time commitment: 3-4 hours/day on weekdays, 4-5 hours on weekends
Core focus areas:
- SQL Mastery (25%): Advanced queries, query optimization, database internals
- Data Modeling (20%): Dimensional modeling, schema design, normalization
- Distributed Systems (20%): Spark, Kafka, data lakes, cloud data services
- Pipeline Design (15%): ETL/ELT patterns, orchestration, data quality
- Coding (15%): Python, Scala/Java, scripting
- Behavioral (5%): Data governance, cross-team collaboration
What makes Data Engineer prep unique: SQL depth matters more here than in any other role. You will face complex SQL problems that test not just correctness but performance optimization and understanding of query execution plans.
Start here: Data Engineer Prep Path
Time Commitment Comparison
Weekly Hours Breakdown
| Path | Weekday Hours | Weekend Hours | Total Weekly | Total Prep Hours |
|---|---|---|---|---|
| MLE (8 wk) | 3-4 hrs/day | 5-6 hrs/day | 25-32 hrs | 200-256 hrs |
| AI Engineer (6 wk) | 3-4 hrs/day | 4-5 hrs/day | 23-30 hrs | 138-180 hrs |
| MLOps (8 wk) | 3-4 hrs/day | 5-6 hrs/day | 25-32 hrs | 200-256 hrs |
| Data Scientist (6 wk) | 2-3 hrs/day | 4-5 hrs/day | 18-25 hrs | 108-150 hrs |
| Research Eng (10 wk) | 4-5 hrs/day | 6-8 hrs/day | 32-41 hrs | 320-410 hrs |
| Data Engineer (6 wk) | 3-4 hrs/day | 4-5 hrs/day | 23-30 hrs | 138-180 hrs |
When to Start Your Prep
:::warning Start Earlier Than You Think Most candidates underestimate prep time by 30-50%. If a path says 8 weeks, budget 10-12 weeks. Life happens -- sick days, work emergencies, mental fatigue. Build buffer into your schedule. :::
Timeline Decision Framework
If You Have Extra Time (12+ Weeks)
Use the first 2-4 weeks before starting your role-specific path to:
- Refresh coding fundamentals -- Work through the essentials in Coding Interviews
- Review your resume and portfolio -- Polish these using Resume and Portfolio
- Start networking -- Connect with people at target companies
- Read broadly -- Skim through the full handbook to identify your weakest areas
If You Are Short on Time (Less Than 6 Weeks)
Follow the "accelerated" variants within each prep path. These cut the lowest-priority activities and focus on the areas that appear most frequently in interviews:
- Always prioritize coding -- It is the most common elimination round
- Focus on your role's highest-weight area -- System design for MLOps, LLMs for AI Engineer, etc.
- Do at least 2 mock interviews -- Even rushed mock interviews reveal blind spots
- Skip depth in your strongest areas -- Do not waste time polishing what you already know
How to Use These Prep Paths
Structure of Each Path
Every prep path in this section follows a consistent structure:
- Role overview and interview format -- What to expect in interviews for this role
- Focus area breakdown -- Percentage allocation across interview dimensions
- Week-by-week schedule -- Detailed daily plans with specific topics and practice problems
- Resource pointers -- Links to relevant handbook chapters and external resources
- Mock interview schedule -- When and how to practice with mock interviews
- Milestone checkpoints -- Self-assessment points to verify you are on track
- Common pitfalls -- Mistakes to avoid for this specific role
How to Follow the Schedule
:::tip The "Good Enough" Principle Do not aim for perfection in every area. Aim for "good enough to pass" in your weaker areas and "exceptional" in your strongest areas. Interviewers hire candidates who are exceptional at something, not candidates who are merely adequate at everything. :::
Daily routine structure:
| Time Block | Activity | Duration |
|---|---|---|
| Morning (before work) | Coding practice | 45-60 min |
| Lunch break | Flashcard review / concept reading | 20-30 min |
| Evening (after work) | Deep study (system design, ML theory, etc.) | 90-120 min |
| Before bed | Review the day's notes | 15-20 min |
Weekend routine structure:
| Time Block | Activity | Duration |
|---|---|---|
| Morning | Mock interview or practice problems | 2-3 hours |
| Afternoon | Deep study or project work | 2-3 hours |
| Evening | Review and plan next week | 1 hour |
Tracking Your Progress
Create a simple spreadsheet or notebook with these columns:
- Date
- Topics covered
- Problems solved (with difficulty level)
- Confidence level (1-5) for each topic
- Notes and questions to revisit
:::note Spaced Repetition Review topics you studied 1 day ago, 3 days ago, 7 days ago, and 14 days ago. This is the most efficient way to retain information. Consider using Anki flashcards for ML concepts and definitions. :::
Cross-Referencing the Handbook
Each prep path references specific chapters of this handbook. Here is a master reference map:
Chapter Reference Map
| Handbook Chapter | MLE | AI Eng | MLOps | DS | Research Eng | Data Eng |
|---|---|---|---|---|---|---|
| AI Career Landscape | Week 1 | Week 1 | Week 1 | Week 1 | Week 1 | Week 1 |
| Interview Process | Week 1 | Week 1 | Week 1 | Week 1 | Week 1 | Week 1 |
| Resume and Portfolio | Week 1 | Week 1 | Week 1 | Week 1 | Week 1 | Week 1 |
| Coding Interviews | Weeks 1-6 | Weeks 1-4 | Weeks 1-4 | Weeks 1-3 | Weeks 1-6 | Weeks 1-4 |
| ML Fundamentals | Weeks 2-5 | Weeks 2-3 | Weeks 3-5 | Weeks 2-4 | Weeks 2-6 | -- |
| Deep Learning | Weeks 3-5 | Weeks 2-4 | Week 4 | Week 4 | Weeks 3-7 | -- |
| LLM Interviews | Week 5 | Weeks 1-5 | Week 5 | -- | Weeks 5-7 | -- |
| ML System Design | Weeks 4-7 | Weeks 3-5 | Weeks 2-7 | Week 5 | Weeks 7-8 | Weeks 3-5 |
| Paper Discussion | Week 6 | Week 4 | -- | -- | Weeks 1-10 | -- |
| Behavioral | Weeks 7-8 | Weeks 5-6 | Weeks 7-8 | Weeks 5-6 | Weeks 9-10 | Weeks 5-6 |
| Take-Home Projects | Week 6 | Week 5 | Week 6 | Week 5 | Week 8 | Week 5 |
| Company Guides | Week 7 | Week 5 | Week 7 | Week 5 | Week 9 | Week 5 |
| Negotiation | Week 8 | Week 6 | Week 8 | Week 6 | Week 10 | Week 6 |
| Curated Lists | All weeks | All weeks | All weeks | All weeks | All weeks | All weeks |
Combining Paths for Hybrid Roles
Many real-world job postings blend responsibilities across roles. Here is how to handle hybrid prep:
MLE + MLOps Hybrid
Some companies expect MLEs to handle their own deployment and monitoring. If you see job postings mentioning "full-stack ML" or "end-to-end ML engineer":
- Follow the MLE path as your primary guide
- Add MLOps system design topics from weeks 4-6 of the MLOps path
- Budget an extra 2 weeks
Data Scientist + MLE Hybrid
Many "data scientist" postings at startups are really MLE roles. Look for keywords like "deploy models" or "production ML":
- Follow the MLE path but add the statistics and A/B testing content from the Data Scientist path
- Reduce system design time by one week
AI Engineer + MLE Hybrid
Some companies want AI Engineers who can also fine-tune and train models:
- Follow the AI Engineer path as primary
- Add ML fundamentals weeks 3-4 from the MLE path
- Budget an extra 2 weeks
:::danger Do Not Try to Prepare for Multiple Roles Simultaneously Pick one primary role and one backup role at most. Spreading yourself across three or more paths guarantees mediocre preparation in all of them. It is better to be exceptional in one area than average in three. :::
Building Your Support System
Study Groups
Interview prep is a marathon, not a sprint. Having a study group dramatically improves accountability and knowledge retention:
- Find 2-3 study partners targeting similar roles
- Meet weekly for mock interviews and problem-solving sessions
- Share resources and discuss tricky concepts
- Hold each other accountable to the weekly schedule
Mock Interview Partners
Aim for this mock interview cadence:
| Prep Week | Mock Interviews | Focus |
|---|---|---|
| Week 1-2 | 0 | Building foundations first |
| Week 3-4 | 1/week | Coding and ML fundamentals |
| Week 5-6 | 2/week | Add system design |
| Week 7+ | 2-3/week | Full simulations including behavioral |
Mental Health During Prep
:::warning Take Care of Yourself Interview prep burnout is real. These schedules are intensive. Build in rest days, maintain exercise, keep social connections, and remember that failing an interview is not failing at life. The skills you build during prep are valuable regardless of any single interview outcome. :::
Non-negotiable self-care during prep:
- At least one full rest day per week (no studying at all)
- 7+ hours of sleep per night
- Regular exercise (even 20-minute walks help cognition)
- Stay connected with friends and family
- Celebrate small wins (completed a week of prep? Treat yourself)
Prep Path Readiness Checklist
Before starting any prep path, make sure you have:
- Chosen your primary target role (and optionally a backup role)
- Set a target interview timeline (when do you want to start interviewing?)
- Blocked time on your calendar for daily study sessions
- Set up your study environment (quiet space, good internet, coding environment)
- Prepared your tracking system (spreadsheet, notebook, or app)
- Found at least one study partner or mock interview buddy
- Read the Interview Process chapter to understand what to expect
- Updated your Resume so you can start applying while you prep
- Bookmarked the Curated Lists for quick resource access
Frequently Asked Questions
"Can I compress an 8-week path into 4 weeks?"
You can, but you will need to double your daily time commitment and accept that your preparation will have gaps. Focus on the highest-weight areas and skip "nice-to-have" topics. See the "accelerated" notes within each path.
"What if I am switching from a non-technical background?"
Add 4-6 weeks of coding fundamentals before starting any path. Focus on Python, data structures, and algorithms. The Coding Interviews chapter is your starting point.
"Should I apply to jobs while prepping or wait until I am ready?"
Start applying by week 3-4 of your prep. Early interviews are learning experiences. You will be nervous, you might fail, but you will learn more from one real interview than from ten mock ones.
"What if I do not know which role I want?"
Read through the AI Career Landscape chapter first. Then look at 20-30 job postings across different roles. Which ones make you excited? Which required skills do you already have? Start with the role that best matches your current skills while still stretching you.
"How do I know if I am ready for interviews?"
Each prep path includes milestone checkpoints. If you can comfortably pass 70% of the checkpoint assessments, you are ready to start interviewing. Remember: you do not need to be 100% ready. Real interviews will continue to sharpen your skills.
Next Steps
Choose your path and begin:
- MLE Prep Path -- 8-week Machine Learning Engineer preparation
- AI Engineer Prep Path -- 6-week AI Engineer preparation
- MLOps Prep Path -- 8-week MLOps Engineer preparation
- Data Scientist Prep Path -- 6-week Data Scientist preparation
- Research Engineer Prep Path -- 10-week Research Engineer preparation
- Data Engineer Prep Path -- 6-week Data Engineer preparation
Pick one path. Commit to the schedule. Trust the process. Your future self will thank you.