Honest Insights

Root causes

• ✗Unrealistic timeline (3 months for complete transition)
• ✗Underestimated math requirements
• ✗No portfolio projects (only tutorials)
• ✗Applied only to senior roles (overestimated readiness)

Lessons learned

• •3 months is not enough for career switch
• •Need real projects, not just courses
• •Start with junior roles, not senior
• •Quality > quantity in applications
• •Math is non-negotiable for ML

What they should have done

• ✓Set 9-12 month timeline
• ✓Build 3-5 solid portfolio projects
• ✓Target junior/mid roles
• ✓Get referrals instead of cold applying
• ✓Take linear algebra seriously

Recovery path

Took 6 month break, then started again with realistic plan - succeeded in month 10

Root causes

• ✗Academic mindset (perfect over done)
• ✗No coding skills (Python basics only)
• ✗Over-focused on theory, ignored engineering
• ✗Poor communication (too technical)

Lessons learned

• •Industry values shipping code, not just papers
• •Need strong software engineering skills
• •Communication matters (explain simply)
• •PhD is not enough - need practical skills

What they should have done

• ✓Learn software engineering fundamentals
• ✓Build deployed projects
• ✓Target research labs (DeepMind) not product companies
• ✓Practice coding interviews

Recovery path

Spent 3 months learning PyTorch and building projects, got offer at research lab

Root causes

• ✗Underestimated technical depth required
• ✗Thought PM role doesn't need coding
• ✗No hands-on AI experience
• ✗Couldn't discuss ML concepts with engineers

Lessons learned

• •AI PM needs deeper technical knowledge than regular PM
• •Must be able to discuss ML concepts
• •Hands-on experience matters
• •Build credibility with engineers

What they should have done

• ✓Take comprehensive ML course
• ✓Learn Python basics and build simple models
• ✓Understand ML metrics and evaluation
• ✓Build no-code AI projects to show understanding

Recovery path

Spent 4 months learning ML basics, built projects with ChatGPT/Claude, succeeded

Root causes

• ✗Bootcamp was too shallow
• ✗Competed with CS grads for junior roles
• ✗Lacked computer science fundamentals
• ✗Projects were all bootcamp assignments (looked same as others)

Lessons learned

• •Bootcamp alone is not enough for ML
• •Need CS fundamentals for interviews
• •Stand out with unique projects
• •Consider starting at smaller companies

What they should have done

• ✓Supplement bootcamp with CS fundamentals
• ✓Build unique projects beyond bootcamp
• ✓Target startups instead of FAANG
• ✓Get internship first to build experience

Recovery path

Worked as data analyst for 1 year, built skills, then transitioned to ML

Root causes

• ✗Didn't update skills (CV background, no LLM)
• ✗Assumed ML experience transfers directly
• ✗Didn't build LLM projects
• ✗Market moved too fast (LLM boom 2023-2026)

Lessons learned

• •Specialization matters - ML is too broad
• •Must stay current with hot areas
• •Build projects in new domain
• •Don't rest on laurels

What they should have done

• ✓Build 2-3 LLM/RAG projects immediately
• ✓Learn prompt engineering and fine-tuning
• ✓Contribute to LLM open source
• ✓Network in LLM community

Recovery path

Eventually built LLM projects, got offer 6 months later