Sarah Sair
Generative AI & Data Engineer · LLM Systems · RAG · Prompt Engineering · GPT-4 · Claude · Gemini · Python · SQL · Chicago, Illinois
Get In Touch
Most AI demos break. I build the systems that don't.
The Difference
With 10+ years of IT infrastructure behind every build, I don't just ask "does this model work?" — I ask whether it's reliable under edge cases, scalable under load, and measurable over time.
What I Deliver
LLM-powered applications, prompt engineering, and AI automation systems that are structured, measurable, and built to perform — not just impress in a sandbox.
What I Build
LLM Systems
End-to-end pipelines using GPT-4, Claude, Gemini, and open-source models — production-ready, not sandbox demos.
RAG Pipelines
Retrieval-Augmented Generation with optimized vector retrieval and context ranking for grounded AI responses.
Prompt Engineering
Zero-shot, few-shot, chain-of-thought, and ReAct workflows focused on consistency and output quality.
AI Automation
LLMs integrated with REST APIs and Python orchestration layers, eliminating repetitive analytical tasks.
Core Tech Stack
AI / GenAI
GPT-4 · Claude · Gemini · RAG · Prompt Engineering · LLM Evaluation
Programming
Python · SQL · scikit-learn · Pandas · NumPy · REST APIs
Data
PostgreSQL · ETL Pipelines · Data Modeling · Star Schema · Data Warehousing
Analytics & Tools
Power BI · KPI Design · Git · Jupyter · DAX · Data Visualization
AI & Data Engineering — Freelance
January 2024 – Present · Chicago, IL
~70%
Less Manual Work
Reduction in manual content generation time via end-to-end LLM pipelines.
~40%
Fewer Hallucinations
Estimated drop in hallucination rates through optimized RAG pipelines.
~60%
Faster Pipelines
Reduction in pipeline turnaround time via AI-to-business workflow automation.
10+
AI Use Cases
Distinct use cases served by advanced prompt workflows with improved consistency.
Also built multilingual prompt chains, structured JSON output systems, model evaluation frameworks, and reusable Python utility libraries — accelerating new project setup across engagements.
Data Engineering & Analytics
April 2025 – Present · Chicago, IL
Pipeline & Modeling
  • Scalable SQL & ETL pipelines for analytics-ready data models
  • Star schema architecture optimized for BI workloads
  • Automated data cleaning, validation & transformation
Analytics & Reporting
  • Power BI dashboards for engagement, retention & performance KPIs
  • Advanced SQL with CTEs, window functions & subqueries — ~50% faster ad-hoc reporting
  • Lifecycle analytics for cohort analysis and churn risk detection
Machine Learning & AI Projects
March 2025 – May 2026 · Independent
End-to-End ML Pipelines
Data ingestion through evaluation using Python & scikit-learn — cutting development cycles from weeks to days.
Model Performance
Classification, regression & clustering models with consistent F1-scores above 0.85 via rigorous cross-validation.
LLM Integration
LLM-based workflows embedded into analytics pipelines, cutting manual analysis effort by an estimated 50%.

A production-grade harness for testing LLM prompts systematically, not by eyeballing output.
  • Four-dimensional weighted scoring (JSON structure, required keys, safety compliance, content rules).
  • Simulated OpenAI provider for continuous testing without API costs. 100% pass rate on test suite.
  • Reduced hallucinations 30–40% through structured prompt optimization.
  • Tech: Python OpenAI API Prompt Engineering LLM Evaluation

A full-stack system for evaluating and refining prompts with structured scoring, iterative comparison, and AI-assisted optimization.
  • Designed a multi-dimensional prompt scoring framework measuring clarity, specificity, robustness, efficiency, and alignment, with hallucination-risk analysis and AI-generated optimization suggestions.
  • Implemented prompt versioning and head-to-head comparison workflows to track iterative improvements and evaluate performance across task types such as RAG, summarization, reasoning, and extraction.
  • Built a full-stack React application integrated with the Anthropic Claude API for real-time prompt evaluation, refinement, and predicted output simulation.
Tech :React18 JavaScript Prompt Engineering LLM workflows Tailwind CSS Anthropic Claude API (Claude Sonnet 4) AI Evaluation Systems
End-to-end analytics pipeline for modeling user performance and engagement.
  • Built leaderboard, streak, and rolling 7-day metrics to track user progress and retention.
  • Developed question difficulty modeling and query optimization to improve performance insights.
Tech: PostgreSQL Advanced SQL Window Functions KPI Modeling
End-to-end machine learning pipeline for predicting telecom customer churn.
  • Feature engineering, model training, and evaluation
  • Business-focused retention insights to support decision-making
Tech: Python Pandas scikit-learn Classification
Fraud detection model designed to address severe class imbalance in credit card transactions.
  • Handled class imbalance with targeted preprocessing and modeling techniques
  • Optimized precision/recall tradeoffs and evaluated performance with F1-score
  • Minimized false positives to improve reliability for real-world detection
Tech: Python Pandas scikit-learn
Credentials & Recognition
Certifications
Azure AI Engineer Associate (AI-102) Practice Exam
Microsoft Azure AI Essentials Professional Certificate
Machine Learning in Telecommunication
Introduction to Prompt Engineering for Generative AI
Honors & Awards
  • TRIO Outstanding Achievement Award
  • Peer Leadership Award
  • Mercer Minority Scholarship
  • CTI Innovative Scholarship
  • NSF Scholarship
Education
B.S. Information Technology · DePaul University
Languages
Urdu (Native) · Punjabi · English · Hindi — all at full professional proficiency
Let's Build Something That Actually Works
Open to AI Engineer, Generative AI / LLM Engineer, Prompt Engineering, and AI & Data Engineering roles. If you're building LLM-powered products, AI workflows, or data-driven AI systems, I'd be glad to connect.
💼 LinkedIn
linkedin.com/in/sarahsair
💻 GitHub
github.com/sarahsair25
Made with