I write about the systems between a company and its customers.

CRM and go-to-market platforms, the data underneath them, and the AI agents now working inside them. I have built these since 2010, and this is where I write about how they're designed.

Now: Technical Lead Manager for Seller Tech at Uber.

Cybage, HCL, ZS 2010 to 2015, Pune Pandora 2014 to 2016 Adobe 2017 LinkedIn 2017 to 2025 Okta 2025 to 2026 Uber 2026 to now GTM systems Data systems LLMs and agents
  1. Cybage, HCL, ZS 2010 to 2015, Pune
  2. Pandora 2014 to 2016
  3. Adobe 2017
  4. LinkedIn 2017 to 2025
  5. Okta 2025 to 2026
  6. Uber 2026 to now
  7. Today the work runs on three lines: GTM systems Data systems LLMs and agents
  1. MCP servers for enterprise systems: the hard parts sit around the protocol

    Architecture notes on MCP servers that connect LLM agents to enterprise systems: tool shape, identity, safe writes, schemas, and evals.

    6 min read Subscribers

  2. Designing a contract data service: exactly-once ingestion and sub-50ms reads

    A systems design write-up of a contract data service: workflow-engine ingestion, Redis locks, MySQL reads under 50 ms, batch mode, and AI-extracted facts.

    6 min read Also on LLMs and agents Subscribers

  3. Using an LLM to clean CRM data: a pipeline pattern that holds up

    An enrichment pipeline for fixing CRM account records with an LLM: structured extraction, per-field confidence, validation, review, and idempotent writes.

    5 min read Also on Data systems Subscribers

Selected work

See all work