Hi, I'm Sanjana —
I turn data into insight and create impact.

Data Scientist | AI Researcher

Sanjana Jairam

About Me

With 2 degrees and nearly 4 years of hands-on experience working with data, I thrive at the intersection of data analysis and engineering. Whether it's building scalable data pipelines, automating workflows with Python, uncovering insights with SQL, or deploying solutions using AWS and Spark — I enjoy making data systems both powerful and elegant.

If you're working on something impactful (or just want to nerd out about LLMs, visualizations, or what makes a clean data pipeline), let's connect.

Storyteller
Coffee Lover
Writer
Visual Learner

Featured Projects

Emi: A Calm Desk Buddy project preview

Emi: A Calm Desk Buddy

A desk companion for students that helps you start tasks, take breaks and focus, designed with neurodivergent users in mind. After auditing a hate-speech filter, I designed Emi’s safety features to avoid its mistakes: swearing and slang are always fine, harsh self-talk gets a gentle reply, and signs of crisis bring in real helplines, never blocking or scolding.

HCI
Human–robot interaction
Accessibility
Safety by design
Who Does a Hate-Speech Filter Fail? project preview

Who Does a Hate-Speech Filter Fail?

My classifier hit 86% accuracy but caught only 2% of hate speech, and flagged harmless African American English six times more often than White-aligned English. A modern transformer model made the same mistakes, more strongly. I traced where the errors come from and tested moderation that keeps people in the loop. Extended from a Social Media Mining course project.

Fairness auditing
NLP
Human–AI moderation

Work Experience

Associate Researcher and Instructor

Indiana University Luddy School of Informatics, Computing and Engineering

Aug 2024 - Present

  • Designed an AI-driven knowledge navigation tool utilizing graph databases like Neo4j, and retrieval-augmented generation (RAG) using LLMs , enhancing information retrieval efficiency.
  • Optimized model scoring techniques, prompt engineering strategies, and Agentic LLM implementation, improving AI-driven decision-making processes.
  • Enhanced chunking and document segmentation methodologies, increasing information retrieval accuracy for the CEDS Planning Process and improving structured document understanding.

Associate Data Scientist

MathCo

Jun 2021 - Jul 2022

  • Built ETL pipelines with Apache Airflow and PySpark, integrated with AWS, GCP, and Azure, reduced manual effort by 40%.
  • Developed supply chain forecasting models that reduced stockouts by 15% and saved $2M annually.
  • Created custom BI dashboards for logistics and operations teams, increasing visibility into performance metrics.
  • Led data cataloging initiatives to improve asset discoverability and standardized metadata practices.
  • Collaborated on implementation of role-based access controls and governance across cloud data environments.

Data Analyst

Pratham Books

Aug 2020 - Jun 2021

  • Improved educational content discoverability by 130% using Bayesian modeling and semantic search optimization.
  • Designed executive dashboards to monitor national UNICEF campaigns reaching 300K+ schools.
  • Consolidated real-time engagement data across MongoDB and Cassandra sources for unified analytics.
  • Conducted A/B testing to optimize outreach strategies, resulting in a 25% increase in user interaction.

Software Developement Engineer

Adobe

Jun 2018 - Jun 2019

  • Built a marketing mix model using Python, SQL, and time-series forecasting (ARIMA, Prophet), leading to $5M in revenue growth.
  • Designed and maintained scalable big data pipelines (Hadoop, Hive, Spark), improving delivery speeds by 30%.
  • Reduced data validation errors by 15% through automated quality checks and pipeline alerts.
  • Produced interactive analytics dashboards for global marketing teams to track campaign ROI.

Skills & Tools

I blend structured logic with creative problem solving.

Languages
Python
Java
Scala
C++
SQL
T-SQL
PSQL
Cloud & Big Data
AWS (S3, Redshift)
GCP (BigQuery, Vertex AI)
Azure (Databricks, Synapse)
Hadoop
Spark
Databases
Snowflake
Cassandra
HBase
MongoDB
PostgreSQL
MySQL
Data Engineering
Airflow
PySpark
ETL automation
CI/CD for data pipelines
Visualization & BI
Tableau
Power BI
Matplotlib
Seaborn
Plotly
Data Modeling
Dimensional modeling
Data marts
Semantic layers
Graph modeling (Neo4j)
Governance & Security
Data cataloging
Access control policies
Hybrid-cloud architecture