Headline

Data Analytics @Trivas.ai | SQL | A/B Testing | DBT | Data Modeling

About

I work at the intersection of data and analytics - I build the data systems that power reporting, then use statistical analysis to turn that data into business decisions. Currently at Trivas.ai, I own data analytics for a multi-tenant e-commerce analytics platform. I designed the platform's underlying data model (Meridian architecture + dbt) with fully isolated business logic per client across 15+ source integrations, and I run A/B tests, cohort and funnel analysis, and significance testing to validate what's actually driving performance. I also own client-facing analytics for enterprise brands and direct a small analytics team. What I focus on: - Statistical analysis -> A/B testing, hypothesis testing, cohort/funnel/attribution modeling, significance testing - Analytics engineering -> dbt data models, SQL ETL pipelines, and multi-tenant architecture (Snowflake, Redshift, PostgreSQL, BigQuery) - BI & reporting -> dashboards and WBR/MBR insights stakeholders act on - Turning messy, multi-source data into reliable, validated reporting Before Trivas.ai, I built analytics from scratch for a 500K+ customer health-tech platform (Viome) and the recommendation engine behind a D2C marketing-intelligence product (Datagets.ai). I'm most interested in Data Analyst / Data Analytics roles where I can dig into the numbers, validate what's real with proper statistical rigor, and turn that into decisions - with the engineering depth to build the pipelines behind it.

Location

Bengaluru

AI stack

antigravity

claude code

claude

gemini

openai

claudecode

Work experience

Data Analyst at Trivas.ai, Aug 2025 - Present · 1 yr 3 mos, Bengaluru, Karnataka, India, → Designed and analyzed A/B tests and cohort-level experiments for Promo Analysis, applying statistical significance testing to validate results, driving a 15% lift in promo-driven conversion. → Built a decision-tree based refund-classification model for AI Analyst, validating classification thresholds against historical refund data through hypothesis testing, eliminating manual review and cutting refund resolution time by 40%. → Engineered client-specific data models and statistical business rules for Wingman, an AI analytics tool, cutting onboarding-to-insight time by 50% across every newly onboarded account. → Own client-facing analytics for enterprise brands, including a $200M+ revenue wellness brand, running A/B tests, cohort analyses, and significance testing on platform audits, informing business recommendations that held 100% client retention across managed accounts. → Designed the underlying data model powering fully isolated, client-specific reporting across 15+ source integrations spanning sales, ads, email, traffic, CS, reviews, subscriptions, and seller platforms, cutting new-client onboarding time by 35% while supporting 6 distinct sub-brands. → Validated and reconciled source-level data across every connected platform through SQL-based ETL checks and variance analysis, powering 100% of accurate client reporting while cutting data-refresh latency by 40%. → Direct a 3-person analytics team, setting day-to-day priorities across client integrations, audits, and reporting deliverables, lifting on-time delivery of reporting commitments to 95%.

Data Analyst at Viome, Oct 2024 - Aug 2025 · 11 mos, Bengaluru, Karnataka, India, Contracted by Invictusdata.ai Data analytics for a 500K+ customer health-tech platform. → Built dashboards from scratch across 8 business domains (Sales, Ads, Traffic, Retention, Subscriptions, Finance, Operations,CX), enabling self-serve reporting for a 500K+ customer base validated to 99+% accuracy against source data. → Applied cohort, funnel, and attribution modeling (Linear, Time Decay) across Sales, Ads, Retention, and Subscription funnels, driving a 12% increase in subscription revenue through data-backed retention and upsell recommendations that helped scale the business. → Presented insights across 40+ WBR/MBR sessions directly to Viome stakeholders, translating analysis into strategic recommendations that shaped quarterly planning. → Led a 4-person migration of Tableau reporting from Redshift to PostgreSQL, delivering on schedule with zero reporting downtime across all live dashboards. → Optimized SQL queries and automated recurring reporting workflows, cutting manual reporting time by 25% while running platform audits that kept monthly financial close on schedule.

Founding Engineer Intern at Datagets, May 2024 - Oct 2024 · 6 mos, Chennai, Tamil Nadu, India, Contracted via Invictusdata.ai Built the recommendation engine behind Datagets.ai's D2C marketing-intelligence platform (contracted via Invictus). → Built Next Best Product, related-product, and cross-sell/upsell recommendation logic for D2C brands, replacing generic one-size-fits-all suggestions and lifting recommendation click-through by 18% without added acquisition spend. → Partnered with the product team to define CTR and conversion-lift metrics for the recommendation engine, driving a 12% conversion lift across live storefronts through iterative tuning → Integrated Shopify data via GraphQL APIs to design ML-ready datasets for predictive modeling, expanding analytics coverage to 10+ additional merchant accounts. → Presented recommendation performance and adoption insights to engineering and product stakeholders, influencing prioritization of 3 roadmap features for the next release cycle

Education

Bachelor of Technology - BTech, Computer Science, DIT University

Public links

portfolio-harshh.vercel.app: https://portfolio-harshh.vercel.app/