SAVAN HIRPARA
I’m a data scientist specializing in Machine Learning, Generative AI, and MLOps, with 6+ years of end-to-end ownership—from model development and experiment tracking to production pipelines that stakeholders can trust.
About
I’m a data scientist specializing in Machine Learning, Generative AI, and MLOps, with 6+ years of end-to-end ownership—from model development and experiment tracking to production pipelines that stakeholders can trust. I enjoy translating business problems into measurable objectives, building robust data foundations, and ensuring models are auditable, compliant, and easy for teams to operationalize.
In my recent work, I’ve built automated analytics pipelines for hyper-personalized marketing, using LLMs to generate and refine campaign content based on conversion outcomes. I also focus heavily on responsible deployment—such as PII detection before messages go out—and I’ve designed reusable data models and statistical evaluation frameworks that reduce time-to-launch while improving confidence across experiments and production use cases.
- 7 years' experience
- Joined January 2026
Industry experience
- Software & Internet
- Professional Services
- Media & Entertainment
- Financial Services
More experts in Ahmedabad, India
Work experience
Owned an automated analytics pipeline that turns raw member behavior into hyper-personalized marketing campaigns, covering audience segmentation through final sends. Used LLMs (ChatGPT, Gemini Pro, Claude) to draft personalized marketing copy and iteratively refine prompts using conversion data. Partnered with credit union stakeholders to scope use cases (e.g., reframing churn as member disengagement). Designed a common data model on top of the data warehouse for reusable measurement across initiatives. Introduced statistical testing across multiple LLMs to validate compliance and sentiment before campaigns launch. Built a PII-detection layer to protect privacy for all AI-generated messages.
Designed and operated cloud-based ML and deep learning pipelines using SQL, Python, and Dataiku (DSS). Managed data in SCD Type 1 and Type 2 tables on Snowflake with stored procedures and dbt. Built modeling suites for member segmentation, attrition prediction (Random Forest), and delinquency risk assessment. Created interactive Power BI and SSRS dashboards to translate model outputs into actionable marketing strategies.
Applied machine learning techniques such as Random Forest, Linear Regression, SVM, K-NN, and K-Means to solve business problems and gain experience taking a model from research question to output.
Education
Hire SAVAN HIRPARA today
Post a job and invite SAVAN, or contact them directly. No hidden commission.