AI Software Engineer
Prachi
Saibewar
Building production AI systems end to end, from real-time ranking and Generative AI workflows at Rokt to agentic RAG platforms of my own.
About
I'm a Software Engineer with 3.5+ years of experience building and owning production AI-enabled, backend, and full-stack systems end to end, from system design and API development through implementation, testing, deployment, observability, and production support.
My focus is the intersection of applied AI and reliable production systems: Generative AI, LLMs, RAG, and agentic workflows, alongside real-time decisioning, distributed systems, evaluation, and performance optimization. I currently work on Rokt's Audience and Bidding platform, and previously built research tooling at the University of Central Florida and enterprise systems at Infosys.
Based in Orlando, FL. Open to relocation.
Experience
PDF Resume ↗-
AI Software Engineer
Built Python/FastAPI backend services for real-time next-best-action ranking, combining first-party, cart, and behavioral signals to score offers, improving NDCG@10 from 0.74 to 0.82 across 50K+ offline evaluation impressions. Integrated identity-resolved customer signals through C#/.NET REST APIs and event-driven data workflows, increasing eligible-signal coverage by 23% while enforcing GDPR/CCPA consent controls. Built Generative AI/LLM experimentation workflows for Rokt's Adaptive Content Engine, automating evaluation and guardrail checks that cut rejected generations from 12.4% to 4.8%. Reduced p95 real-time decisioning latency from 185ms to 132ms through Redis caching and asynchronous signal retrieval, and built a React/TypeScript decision-debugging console that cut investigation time by 30%.
-
Research Engineer
Modernized the Chartalist research platform, migrating legacy PHP functionality to reusable React/TypeScript components to improve maintainability and iterative development. Implemented and optimized AlphaCore graph-analysis algorithms in Python and NetworkX, improving experimental runtime by ~30% through vectorized operations and caching. Engineered feature-engineering pipelines for Bitcoin chainlet-orbit analysis, transforming raw transaction structures into 48-dimensional daily feature vectors for downstream graph analytics and ML experiments.
-
Senior Systems Engineer (Software Engineering)
Developed enterprise full-stack financial applications using React, TypeScript, Node.js/Express, and Oracle SQL/PLSQL, supporting production workflows processing 3,000+ daily customer and account transactions. Improved API response times by 25% through query and index optimization, and built REST-based healthcare integrations supporting 5,000+ daily transactions while reducing data-quality defects by 20%. Expanded Jest and Postman test coverage to cut regression defects by 18%, and reduced production issue-resolution time by 40% through systematic root-cause analysis.
Education
M.S. Computer Science, GPA 3.9
B.E. Computer Science
Skills
- Languages & Development
- Python, C#, TypeScript, JavaScript, SQL, FastAPI, .NET, ASP.NET Core, Node.js, Express.js, React, REST APIs, Microservices
- AI / LLM
- Generative AI, LLMs, RAG, LangGraph, Embeddings, Vector Search, Structured Tool Calling, Prompt Engineering, LLM Evaluation, Guardrails, Human-in-the-Loop
- Data & Cloud
- PostgreSQL, pgvector, Redis, SQL Server, Oracle SQL/PLSQL, AWS (S3, SQS, Batch), Event-Driven Architecture
- DevOps & Observability
- Docker, CI/CD, Jest, Playwright, Postman, OpenTelemetry, Prometheus, PromQL, Chronosphere
- AI-Assisted Development
- Cursor, Claude Code, GitHub Copilot
Get in Touch
Open to AI Software Engineer opportunities. Email me directly or use the form.