Skip to main content
← back

Lead, Full-Stack Team · 2 years · 6 engineers

Building ALPHA10X: an AI-Native Decision Intelligence Platform

As lead of the full-stack team, I took ALPHA10X from a fragile, hard-to-onboard React codebase to a production-ready Angular, NestJS and Neo4j platform — leading a six-person team through a two-year build of an AI-native decision intelligence platform for private markets.

50M+
companies in the knowledge graph
~1B
entrepreneurs, experts & innovators
6
engineers led
2 yrs
from rebuild to production
Neo4j
large-scale knowledge graph
Production-ready
outcome

Context

ALPHA10X is an AI-native decision intelligence platform for private markets. It maps over 50 million companies, tens of thousands of technology trends, and close to a billion entrepreneurs, experts and innovators into a single knowledge graph — connecting promising startups with the investors and innovators who can help them grow.

The challenge

Engineering velocity under constant team churn

I inherited a messy React application that was difficult to onboard onto, while the team itself was unstable — members joining and leaving frequently. With that much churn, convention beats configuration: the codebase had to stay legible no matter who was working in it.

Query performance on a massive knowledge graph

ALPHA10X's value lives in a very large Neo4j knowledge graph. For an interactive product, graph traversals across tens of millions of nodes and their relationships had to stay fast — performance was a first-class requirement, not an afterthought.

My approach

Re-architected to Angular for durability

I migrated the front end to the latest Angular. Its opinionated, batteries-included structure — modules, dependency injection, a standard CLI, enforced patterns — means every part of the codebase looks the same. A new engineer lands in familiar territory on day one, which cut onboarding time and kept a rotating team productive and the project on track.

Made Neo4j query performance a first-class concern

I indexed the anchor nodes that traversals start from so every query begins with an indexed seek rather than a scan; bounded traversals by capping path depth and expanding only the needed relationship types; profiled and parameterized hot Cypher queries to remove cartesian products and reuse cached query plans; returned projections over full nodes with keyset pagination for large result sets; and cached expensive, frequently-requested traversals at the NestJS API layer.

Led the team and the delivery

I owned the roadmap, ran sprints and releases, mentored junior developers, and hired into the team — keeping a six-person group shipping coherently across a two-year build.

Outcome

  • The redesign and new architecture took the platform from fragile to production-ready.
  • A maintainable, conventional codebase absorbed team churn without losing momentum.
  • The platform went on to attract customers on the back of the rebuilt foundation.

Stack

  • Angular
  • NestJS
  • Neo4j
  • Azure

Need an engineer who can lead a build like this end-to-end?

Get in touch →