SpotDraft Multi-Agent Legal AI System
Designed and led the full architecture of a multi-agent AI platform for in-house legal teams, owning routing, planning, agent tooling, and infrastructure.

Pranav Pandey
Senior Applied AI Engineer at SpotDraft





From their time as

Senior Applied AI Engineer
SpotDraft β’ 2025
Overview
Pranav led the design and build of SpotDraft's multi-agent AI system for in-house legal teams. The system served hundreds of enterprise clients and required decisions across every layer of the stack: which agents were needed, how they interacted, what sub-agents and tools each required, how context was managed within each agent, and how the overall infrastructure would scale.
The Story
Pranav led the design and build of SpotDraft's multi-agent AI system for in-house legal teams. The system served hundreds of enterprise clients and required decisions across every layer of the stack: which agents were needed, how they interacted, what sub-agents and tools each required, how context was managed within each agent, and how the overall infrastructure would scale.
He owned the routing and planning architecture, designed the agent interaction model, and built in capabilities including code execution, long-term memory, file storage, and document access. Every structural decision about how the system would behave in production was his to make and defend.
The system was built from scratch in a fast-moving environment where competitive pressure from a major foundation model provider created urgency. Pranav made the call to invest in evaluation infrastructure alongside the core system, arguing that measurable product depth would matter more than feature velocity in the legal AI market.
