Open to opportunities

Building AI agents and production systems that move the needle.

Currently at American Express Canada — shipping LLM-based agents, RAG pipelines, and high-throughput backend services at scale. 3.5 years building systems that handle millions of users and eliminate weeks of manual work.

// Experience

Where I’ve shipped

American Express Canada

Software Engineer II

Oct 2024 – Present · Hybrid, Toronto

  • Built a conversational AI bot at an org-wide hackathon (team of 2–3) using Amex’s internal LLM platform — an iterative agentic loop with three intents (QUERY, REASONING, RESPONSE) that lets business and portfolio managers ask natural language questions against prebuilt STAR schema views; the agent combines NL-to-SQL generation and predefined query tools to surface business growth insights on demand.
  • Built a net-new Third Party Onboarding Portal replacing a fully manual email-and-spreadsheet process that had accumulated a backlog of 200 cases and completed only 10–15 per year — the automated platform handles 500+ cases annually (33× throughput increase).
  • Optimized legacy QR code generation APIs from ~2s to ~200ms (90% latency reduction) by introducing database indexes, rewriting slow queries, and offloading non-critical blocking business logic to a background queue.

Glitchy

Senior Software Engineer

May 2024 – Oct 2024 · Remote, Los Angeles

  • Orchestrated a 6-member engineering team, achieving a 100% on-time delivery rate for all product milestones.
  • Instituted rigorous GitHub Actions checks and comprehensive code reviews, leading to a 30% reduction in post-release defects.
  • Architected scalable solutions for new features, supporting a 60% increase in user traffic without service degradation.

Viral Nation

Software Engineer

June 2023 – May 2024 · Hybrid, Toronto

  • Engineered and optimized GraphQL APIs and backend services, achieving a 50% improvement in data retrieval speeds.
  • Architected a full-stack system integrating BFF, middleware, orchestration, and backend layers, bolstering scalability by 45% and reducing PostgreSQL query times by 80% through strategic Redis caching.

Nanoleaf

Full Stack Developer Intern

Sept 2022 – Dec 2022 · Hybrid, Toronto

  • Developed an ETL pipeline integrating diverse data sources with Google BigQuery, reducing manual data handling by 50%.
// Applied AI

Agentic systems in production

Not side projects — shipped AI work at one of Canada’s largest financial institutions and open-sourced a working orchestration system.

Open Source · github.com/rutvij26/chess-context

Chess Context MCP

Semantic chess intelligence MCP server for Claude. Provides position analysis, game review, and player scouting through the Model Context Protocol — giving LLMs structured access to chess knowledge.

MCPTypeScriptClaudeChess
View on GitHub →
Org-wide Hackathon · Amex Canada · Team of 2–3

Conversational AI Bot — Business Intelligence

Built an iterative agentic loop with three intents — QUERY, REASONING, RESPONSE — that lets business and portfolio managers ask natural language questions against prebuilt STAR schema views. The agent combines NL-to-SQL generation with predefined query tools to surface business growth insights on demand, using Amex’s internal LLM platform.

Agentic loopNL-to-SQLTool callingLLM orchestration
Open Source · github.com/rutvij26/incident-orchestration-agent

Incident Orchestration Agent

Multi-agent system that coordinates LLM-powered agents to investigate production incidents. Decomposes incidents into structured investigation tasks, analyzes system signals, and generates actionable remediation suggestions. A working demonstration of multi-agent orchestration, tool-call patterns, and retrieval-driven reasoning at production quality.

Multi-agentOrchestrationTool callingOpen source
View on GitHub →
// Projects

Selected work

TypeScriptMulti-agentLLMOrchestration

Incident Orchestration Agent

AI-driven incident orchestration system that coordinates multiple LLM-powered agents to investigate production incidents. Decomposes incidents into structured investigation tasks, analyzes system signals, and generates actionable remediation suggestions.

TypeScriptMCPChessClaude

Chess Context MCP

Semantic chess intelligence MCP server for Claude. Provides position analysis, game review, and player scouting through the Model Context Protocol.

TypeScriptGitHub ActionsAICode Review

Sentinel

Production-ready GitHub Action for AI-powered PR reviews. Automatically analyzes pull requests using OpenAI or Gemini and posts structured review comments.

Next.js 14TypeScriptTailwind

Portfolio v3

This site. Built with Next.js 14 App Router, TypeScript, and Tailwind CSS.

// Skills

Tech I work with

AI / Agents

LLM OrchestrationRAG PipelinesAgentic LoopsNL-to-SQLTool CallingPrompt Engineering

Languages

TypeScriptJavaScriptGoC#SQLRubyBash

Backend

Node.jsAdonisJSExpress.jsFastifyRESTful APIsGraphQLWebSockets

Frontend

React.jsNext.jsTailwind CSSReduxZustand

Databases

PostgreSQLMySQLMongoDBRedisCassandraBigQuery

Cloud & DevOps

AWS (EC2, S3, RDS)GCP (BigQuery, Firebase)DockerGitHub ActionsCI/CD
// Contact

Let’s build something

I’m currently open to Software Engineer II / III roles in Canada — ideally at the intersection of applied AI and backend engineering. If you’re building something interesting, I’d love to hear about it.

Send me a message →