PhD chemist → applied AI & data analysis. I evaluate and build with large language models, design prompting and AI workflows, and automate data work in Python.
🔬 9+ years turning complex, multi-source data into clear conclusions. 🤖 Now focused on applied AI: LLM evaluation, prompt engineering, MCP workflows, Python automation.
- llm-cv-tailoring-engine — a system-prompt engine that turns a job description + structured evidence into a tailored, truthful one-page CV with a full numerical audit. Truth-as-constraint design, non-compensatory fit scoring, 17-gate validation.
- ai-enablement-toolkit — a practical toolkit for reviewing and improving Custom GPT prompts: prompt cards, test cases, guardrails, and failure-pattern logs for human-in-the-loop validation.
- math-arena — a browser-based mental math trainer for consulting/quant-trading interview prep: adaptive difficulty, timed drills, 40+ pattern-specific task generators.
- Fortune_telling_bot & Luck_coach_bot — two Telegram bot prototypes: one LLM-powered (prompt design, i18n, voice input), one deterministic (SQLite stats, data persistence).
Python · LLM evaluation · Prompt engineering · Custom GPT / GPT Builder · Telegram Bot API · LLM APIs (DeepSeek/OpenAI) ·
MCP servers · Claude Code · SQLite / relational data · Whisper / voice-to-text · Bash · Excel/VBA automation · data cleaning & analysis
SQL · Power BI (Microsoft PL-300 track) — building toward BI / data-analytics reporting.
- LinkedIn: https://www.linkedin.com/in/smirnovilia/
- Google Scholar: https://scholar.google.com/citations?user=_plFef8AAAAJ&hl=en
- YouTube: https://www.youtube.com/@Ilia_Smirnov/videos