Как я за 3 месяца зашёл в ML (без платных курсов)

Summary

Nikita describes how he went from basic Python knowledge and limited machine-learning experience to interview preparation in three months, followed by about two months of interviewing. He credits his progress less to special courses than to a temporary, structured routine, consistent study, and guidance from a mentor.

Learning approach

• Follow a roadmap rather than trying to study everything. Nikita covered Python and algorithms, data analysis, core math, classical machine learning, deep learning and NLP, LLM engineering and production, plus some SQL and statistics.
• Prioritize material relevant to the roles you want. Learn core concepts deeply; skim or postpone specialized topics that are unlikely to matter.
• When a topic is confusing, build a high-level understanding first, then break it into specific points to learn. Use concrete examples and analogies to make abstract ideas clearer.
• Don’t get stuck trying to master every mathematical derivation. Focus on the math needed to understand the models and do the work.
• Use AI assistants to explain concepts, check your understanding, and generate examples—not as the sole authority on interview requirements or what matters in a job. A mentor or experienced practitioner can help filter and prioritize topics.
• Expect gaps and forgetting. Revisit important, frequently tested topics over time rather than insisting on perfect understanding immediately.

Productivity and self-care strategies

• Treat motivation as unreliable; use routines and discipline to begin even when you don’t feel inspired. Nikita describes his intensive study schedule as a temporary push, not a sustainable lifelong pace.
• Set a study-time target—he aimed for about five hours a day—rather than a vague goal like “finish five topics.”
• Track focused study time with a timer. He used a 40-minute work, 20-minute break cycle to measure actual effort and limit self-deception about how long he studied.
• Use a calendar to plan and review the day. Seeing study blocks alongside sleep and other activities can reveal where time is going and provide a visual accountability cue.
• Build small habits that create momentum. His examples include five minutes of morning exercise, logging expenses, and listening to an English podcast during his morning routine.
• If you miss a day or feel stuck, don’t turn a setback into an extended slide into distraction. He recommends taking a walk, spending time outdoors, or resting, then returning to study after a short break.
• Be realistic about progress: it may not feel visible day to day. Recognize the value of showing up consistently, and revisit difficult material instead of assuming that confusion means you can’t learn it.
• Nikita says a bit of self-directed frustration helped him restart, but he also cautions against harshly blaming yourself. This is his personal tactic, not a universal recommendation.

Presenters and sources mentioned

• Presenter: Nikita, an AI engineer.
• James Clear: Atomic Habits.
• Other sources or tools discussed: An unnamed mentor and learning roadmap; ChatGPT and other AI assistants; Google Calendar; Google Sheets; a Pomodoro-style timer.