I am relatively agnostic to industry, but generally interested in B2B SaaS with a strong founding team preferably with a history of being in startup land
I want to explore how advancements an LLM, memory systems and agentic workflows can be used to improve any existing problem or emerging problem in different industries
Architected and built the system enabling security practitioners to schedule real-time and historical analysis of security data in Snowflake for Okta. This system was responsible for running all of Okta's internal security detections.
March 2025 - Ongoing
- Led engineering to validate building a tech company on top of a media agency
- Conducted customer interviews, market research, and development cost modelling
- Facilitated product ideation sessions focused on LLM applications for efficiency and client outcomes
- Built an LLM, RAG-powered workflow for Google Ads optimization that outperformed human accuracy
- Implemented ETL from ad platforms to Snowflake with evaluation pipeline for LLM performance
- Created classification workflow using Snowflake, Pinecone and Python statistics libraries
- Developed React JS frontend for human-in-the-loop triage workflow
June 2022 - March 2025
- Architected and built a system enabling security practitioners to schedule real-time and historical analysis of security data in Snowflake; ran ~10,000 security detections daily across the data platform
- Set up CI/CD pipeline using CircleCI and Terraform for infrastructure management
- Developed Python codebase on a serverless, event-driven architecture (EventBridge, SQS, Lambda, Snowflake)
- Accelerated detection development with open source libraries (SQLGlot) and robust testing practices
- Contributed to a custom ETL pipeline (Lambda, DuckDB, Snowflake) processing TBs of security data daily
- Built security automations integrating third-party APIs and Slack bots to streamline investigation activity for analysts
January 2022 - June 2022
- Threat-modelled and wrote security detections on top of Snowflake data to identify suspicious activity across SaaS applications
- Investigated security incidents triggered by these detections and external events
- Built security automation to improve daily reporting and detection-writing
October 2019 - December 2021
- Built services that classified 10 million inbound emails daily using Python on top of S3, Kinesis, SQS, and Postgres
- Owned and built Terraform and CI/CD pipelines for all services
- Developed Python services for data aggregation and NodeJS APIs for the React frontend
- Optimised multi-tenanted Postgres database, managing migrations and roll-outs
- Co-built an internal authentication framework and systems for secure multi-tenanted data access in AWS
- Created a framework for rapid demo account setup (under an hour) for different client profiles
- Shipped features from discovery through testing to implementation and roll-out
- Conducted security research in spare time, identifying various vulnerability classes in the product
