Staff Solutions Architect · Brooklyn, NY

Taking products from first customer to real scale.

I’m Akash. For eight years I’ve been the engineer customers talk to: I design the architecture, write the code when it needs writing, and build the AI tooling that lets a product scale once engineers are out of the loop. Right now that’s signal infrastructure and agentic commerce at Meta.

$1B in ad spend through Meta Signals Gateway within its first year
30 days → 1 click advertiser onboarding, rebuilt as self-service
+40% spend from advertisers who adopted the signal stack, A/B tested
300 people in Meta’s Solutions org running on a GTM tool I built

Selected work

What I’ve shipped

Meta · Signal infrastructure

Meta Signals Gateway

First-party signal infrastructure that collects, dedupes, and routes conversion data from web, app, CRM, and point-of-sale to Meta, BigQuery, and other destinations. I set the technical roadmap, made the call to go serverless, and owned it from pre-GA to $1B in ad spend in under a year.

Meta Superintelligence Labs · Agentic commerce

Muse partner integrations

Leading the enterprise partner and connector integrations that bring an existing portfolio of major retail and commerce brands onto Muse, Meta Superintelligence Labs' personal AI agent, to power agentic shopping across their catalogs.

Meta · Applied AI

Signal diagnostics MCP

An MCP server in Python and TypeScript that audits advertiser signal health from the account down to a single ad set, then routes each issue to the people who can fix it. Up to a week of manual investigation per advertiser became one command, now used by 100 people a week across Meta’s top 300 advertisers.

Meta · Product engineering

Conversions API Gateway, multi‑account

Designed in Figma and built the React front end for the multi-tenant flow that lets agencies run many client accounts from one configuration, partnering with backend engineers on tenant isolation.

Meta · GTM tooling

Account prioritization

A nightly pipeline and dashboard (TypeScript, Python, SQL, Claude) that ranks accounts by untapped headroom rather than raw size, with collision-free claiming so reps never overlap. Started in North America and now runs globally.

CoEnterprise · Financial services

Equifax on Google Cloud

Led the Python automation workstream for moving sensitive credit-data workflows to GCP. Ran legacy and new processes in parallel, diffed every output automatically, and cut over after a clean validation week.

Projects

Outside of work

Experience

Where I’ve been

  1. 2021 – Present

    Meta

    Staff Solutions Architect

    Signals Gateway, app signals and Apple privacy strategy, AI tooling for the Solutions org, and now Muse partner integrations. Pod lead for a team of eight, owner of the Solutions Architect hiring pipeline, and Meta’s technical representative to the IAB Tech Lab CAPI working group.

  2. 2018 – 2021

    CoEnterprise

    Software Engineer

    Data pipelines and automation for financial-services clients on AWS and GCP, including Equifax and Wells Fargo, where the tooling grew billed project revenue by 60%.

  3. 2017

    New York University

    B.S. Integrated Digital Media

    Tandon School of Engineering, software engineering track.

Languages
Python, TypeScript, SQL, Java
AI
Claude API, MCP, agentic workflows, LangChain, PyTorch
Cloud & data
GCP, AWS, BigQuery, Snowflake, Spark, Airflow
Infrastructure
Docker, Kubernetes, Terraform, serverless, CI/CD

Contact

Let’s talk.