Software Engineer focused on backend systems, distributed APIs, agentic tools, and LLM-powered applications.
I build production-oriented AI software systems with Python, Flask, FastAPI, AWS, semantic search, browser automation, and LLM orchestration.
- Build backend services for LLM-powered applications and AI products
- Design RESTful APIs, async task pipelines, semantic search, and RAG systems
- Optimize LLM workloads for cost, latency, token usage, and retrieval quality
- Develop agentic browser automation tools using Chrome DevTools Protocol
- Ship and monitor cloud-based services on AWS
🔹 Currents — Intelligent Social Search Engine · currents.one
A search engine that surfaces how online communities discuss products, brands, and topics, turning scattered social content into structured insight.
- Built and extended backend services for LLM-powered semantic search using FastAPI and AWS
- Designed asynchronous task-processing pipelines with SQS, Lambda, S3, and Boto3
- Refactored keyword-extraction logic, reducing third-party LLM API calls by 66%
- Migrated vector storage from Milvus to AWS S3 Vector DB, reducing infrastructure cost by $400/month while maintaining retrieval quality
- Added CI/CD and production observability with GitHub Actions and CloudWatch
Tech: Python · FastAPI · RAG · Embeddings · MySQL · AWS EC2 · SQS · Lambda · Boto3 · S3 Vector · DynamoDB · CloudWatch
🔹 Extends — Autonomous Web Agent · extends.one
An LLM-powered Chrome extension that helps users automate multi-step web workflows through natural language.
- Built browser automation features using Chrome DevTools Protocol, Chrome Extension APIs, and LLM-driven task execution
- Developed orchestration components for planning, tool execution, recursive summarization, semantic filtering, and PDF parsing
- Improved context handling to reduce token usage and improve response relevance
- Integrated AWS AgentCore Memory for cross-session personalization
- Evaluated the agent on WebVoyager-style browser automation tasks across dynamic websites
Tech: TypeScript · React · Next.js · Flask · Chrome DevTools Protocol · OpenAI SDK · Gemini SDK · AWS
🔗 Eval repo: Virtu-llc/web_agent_eval
Backend & Infrastructure: Python · FastAPI · Flask · REST APIs · Redis · MySQL · PostgreSQL · Docker · Linux
AI & LLM Systems: OpenAI SDK · Gemini SDK · RAG · Embeddings · Agent Orchestration · Context Management · Prompt Engineering
Cloud & DevOps: AWS EC2 · S3 · SQS · Lambda · DynamoDB · CloudWatch · GitHub Actions
Vector & Retrieval: Milvus · AWS S3 Vector DB · AWS AgentCore Memory · Semantic Search
Frontend & Web Tools: TypeScript · React · Next.js · Tailwind · Chrome Extension APIs · Chrome DevTools Protocol