Software Engineer · Backend • Full-Stack • Applied AI · Orlando, FL
I'm a software engineer with 3+ years building production APIs, event-driven integrations, and the products on top of them, mainly in C#/.NET, Python/FastAPI, React/TypeScript, SQL, Kafka, and Redis.
Most recently at Rokt (mParticle) I worked on a customer-data platform spanning 300+ integrations: Customer 360 identity resolution, real-time audience activation, consent-aware routing, and API performance. Before that I built Python/FastAPI RAG and evaluation tooling as a Research Engineer at the University of Central Florida, and C#/.NET services and Kafka integrations at Infosys.
Applied AI is where I go deeper: retrieval, LLM evaluation, and agentic workflows, built and measured like any other backend system.
prachisaibewar.com · LinkedIn · prachi.saibewar.cs@gmail.com
IncidentIQ: agentic AI platform for production incident support. A LangGraph agent with hybrid retrieval (pgvector + Postgres full-text), read-only diagnostic tools, persisted investigation state, and approval-gated write actions, all on a FastAPI backend. Its evaluation harness publishes real results, including ones still below target. Retrieval measures Recall@5 0.80 and MRR 0.92. The approval API, frontend, and tracing are in progress.
Python FastAPI LangGraph PostgreSQL pgvector Docker
Chartalist.org: catalogue of ML-ready blockchain graph datasets from the NeurIPS 2022 Chartalist paper. I rebuilt the platform off legacy PHP into a React/TypeScript application.
React TypeScript Research Platform
Geo-Enabled POI API: Points-of-Interest REST API with spatial indexing, optimized radius queries, tests, benchmarks, and an interactive map UI.
Django 5 GeoDjango PostGIS
FinAgent Pro: multi-agent investment research with market, fundamental, and technical analysis agents producing structured research reports.
Python Multi-Agent LLM
| Contribution | Status |
|---|---|
| FDataLab/Chainlet-Orbits#1, transaction ETL tooling for the Chainlet Orbits (KDD 2025) codebase | Merged |
| networkx/networkx#7974, adding the AlphaCore data-depth core decomposition algorithm | In review |
| friedhelmvictor/alphacore#3, refactor of the reference Python & R implementations | In review |
Languages: Python, C#, TypeScript, JavaScript, SQL
Backend & distributed systems: FastAPI, .NET/ASP.NET Core, Node.js, REST APIs, Microservices, Kafka, Redis, Event-Driven Architecture
Frontend: React, Next.js, Redux, HTML/CSS
Data & cloud: PostgreSQL, MongoDB, AWS, Docker, Kubernetes, CI/CD, GitHub Actions
Testing & observability: Jest, Playwright, Postman, OpenTelemetry, Prometheus/PromQL, Chronosphere
Applied AI: LLMs, RAG, AI Agents, LangGraph, LangChain, pgvector, Vector Search, LLM Evaluation, Structured Outputs, Guardrails, Human-in-the-Loop