Anuroop Jajoba.
Software engineer building production backend systems and agentic AI. Currently building @PaySplit, with an MS in Information Technology from the University of New Hampshire.
Selected work ↓0
Projects shipped
0%
Query latency cut
0M+
Records processed
0%
Best F1 score
Selected work.
/ projectsMore experiments.
/ archiveExperience.
/ experiencePart-time
Software Engineer · Mercor
Evaluate and annotate real-world TypeScript and backend tasks for frontier models, scoring correctness and spec clarity. Identify edge cases and failure modes in model output, build reproducible minimal repros, and feed fixes back into evaluation datasets and reviewer rubrics.
Part-time
AI Expert Contributor, Full-Stack Coding · Snorkel AI
Author and review full-stack Python and TypeScript changes — including Node.js service endpoints — with a focus on correctness and secure defaults. Build internal eval harnesses, unit tests, and scoring rubrics for LLM workflows, turning messy real-world processes into clear acceptance criteria.
Part-time
AI Evaluation Engineer · Handshake AI, Project Dynamo
Design Python and TypeScript tasks with automated validation and rubric-based scoring for agent evaluation — promoted from Contributor to Reviewer. Built a stateful data-validation task that held reference agents to 0/5 at pass@5 while staying provably solvable, and found a correctness vulnerability where an agent scored 120/120 without solving the task, shipping a proof-of-concept and driving a fix with regression coverage.
Full-Stack SWE & Founding Engineer · PaySplit
Building full-stack features end to end — React TypeScript frontend, FastAPI and Spring Boot microservices, PostgreSQL schema design, and Redis caching on AWS ECS at 99.9% uptime. Shipped LangGraph multi-agent fraud detection with the Claude API and an XGBoost model at 94% accuracy on sensitive financial data.
Research Intern · University of New Hampshire
Built an AI workflow that automates hours of manual document review — ChromaDB RAG pipeline with the Claude API at 91.6% F1 across 161 real documents, deployed to AWS via Jenkins CI/CD with 100% uptime. Worked directly with subject-matter experts to turn ambiguous requirements into a production system.
Software Engineer · Infosys
Built scalable backend services and REST APIs in Python and Java Spring Boot — 500K+ daily records across 8+ production microservices on AWS and Azure. Optimized PostgreSQL to cut query latency 40% and automated CI/CD, cutting deployment failures 30%.
About.
/ about
Anuroop Jajoba · software engineer
I'm a software engineer who cares about the whole system — from PostgreSQL schema design and Kafka event streams to the agent pipelines that sit on top of them.
I hold an MS in Information Technology from the University of New Hampshire (May 2026), and before that spent a year and a half shipping production microservices at Infosys. These days I split my time between PaySplit, where I'm the founding engineer, and AI evaluation work at Handshake AI, Snorkel AI, and Mercor — designing the tasks, rubrics, and harnesses that measure how well frontier models actually write software.
AWS Cloud Practitioner · Neural Networks & Deep Learning (DeepLearning.AI) · Building AI Agents (IBM)
Languages
Python, TypeScript, JavaScript, Java, Go, SQL, Swift
Frontend & Mobile
React, Next.js, Tailwind CSS, SwiftUI, iOS
Backend & APIs
FastAPI, Spring Boot, Node.js, REST, Microservices, Kafka
LLM & AI
LangGraph, LangChain, Claude API, OpenAI API, RAG, ChromaDB, FHIR R4, LLM evals & agent benchmarks
Cloud & DevOps
AWS, Azure, Docker, Kubernetes, PostgreSQL, Redis, GitHub Actions, Jenkins
ML & On-device
PyTorch, CoreML, MediaPipe, scikit-learn, XGBoost, MLflow