At Highspring, I contribute to a Google project evaluating AI Mode query outputs for user harm, neutrality, and preachiness. I also built a Python analytics pipeline to assess notable-entity content, identify missing sections, and reduce manual analyst effort by 87%.
I use Gemini through Google Colab with prompt engineering and few-shot prompting to support consistent query classifications. I developed PLX dashboards that track query volumes, evaluation activity, and manual versus LLM processing time.
At NITS Solution, I engineered a Python-based fuzzy matching solution with Pandas and RapidFuzz for automotive records. It automated matching across 10K+ records and reduced manual SQL analysis workload by 80%.
At NEC Corporation India Pvt. Ltd., I managed and validated 16 enterprise datasets containing 10M+ records. I designed Power BI dashboards for crime trends and KPIs and automated Excel reporting workflows, reducing report preparation time by 30%.
