yashika pal

Through my learning period I've been building AI SaaS apps and tools hands-on, mostly working on OCR and data-based agentic pipelines, plus AI security and governance in my current workplace. This time really pushed me deeper into AI agents and automation. I built PharmaLens, an AI tool using hybrid OCR pipelines to read medicines and prescriptions, along with a small attendance tracker and a supply-and-demand chain AI insights dashboard. I've also worked closely with Purview and AI governance/security over the last few months, which taught me that building the agent is only half the job — making it safe and compliant is the other half.

Along the way I got into automation pipelines and vibe-coding tools like Claude Code, and learned that coding with these tools is fast, but only when you code with clear boundaries and really detailed prompts. QA testing after every shipped feature also matters way more with vibe-coded apps than with manually written code. That's how I've picked up most of my AI/ML skills — through real, hands-on industry work, not just theory.