Experience

Sole developer of the flagship AI SOC/SOAR platform

designed and shipped alert ingestion & triage, investigations & case management, SOAR playbooks with approval gates, and compliance reporting end-to-end, at team velocity via AI-assisted development

AI task & agent engine, 43 task types

incident response, compliance audits, fraud detection, and code review across in-process, Dockerized (warm container pool), and remote-runner modes, with approval gates, heartbeats, and per-run AI decision logs (tokens, cost, latency)

105 integrations across 34 categories

SIEM, EDR, ticketing, IAM, threat intel, and AML/fraud, with credential rotation and an internal JWT bridge so task code never holds customer credentials

Security & quality hardening

Casbin RBAC, JWT + MFA, tamper-evident audit logs, and replay-protected HMAC-signed webhooks; air-gapped deployments covered by 150+ backend and Playwright E2E tests

Billion-scale data engine

architected on ClickHouse, Dask, and Prefect to process 1B+ telecom & telemetry records, cutting query latency from minutes to milliseconds

OSINT & vector search system

integrated Telegram/YouTube APIs with NLP pipelines for intelligence gathering

TypeScript agent SDK

bridged backend AI agents with frontend apps, enabling 200+ MCP tool integrations

LLM fine-tuning datasets

delivered domain-specific training and evaluation sets that improved model accuracy

ETL automation

built Python/SQL pipelines with AWS-backed storage workflows, cutting manual preparation time

Selected from 25,000+ applicants for Amazon's intensive machine learning program in consecutive years

Covered supervised learning, deep learning, NLP, and probabilistic graphical models under Amazon scientists