Edison Scientific Tech Stack

AI platform for scientific R&D workflows and discovery acceleration

Artificial Intelligence San Francisco, California 51–200 employees Founded 2025 Privately Held

Edison Scientific built an AI scientist (Kosmos) to integrate across the full R&D lifecycle—discovery through regulatory approval. The stack is modern and distributed: Python + PyTorch + JAX for model work, Node.js + FastAPI + GraphQL for APIs, React + Next.js for UI, and multi-cloud infrastructure (AWS, GCP, Azure) with Kubernetes orchestration. The hiring shape is engineering-heavy (18 of 28 roles) with principal and senior emphasis, and active pain points around Kubernetes scaling and resilient scheduling for long-running workloads signal they're scaling compute-intensive experimental pipelines in production.

Tech Stack 35 technologies

Core StackPython TypeScript Ashby Node.js FastAPI GraphQL React Next.js PostgreSQL MySQL MongoDB AWS Docker Kubernetes Terraform Pulumi Prometheus Grafana Datadog PyTorch JavaScript GCP Azure Jax Ray Megatron-LM WebSockets D3.js Deck.gl Three.js+5 more

What Edison Scientific Is Building

◆Challenges

Slow scientific discovery
Resilient scheduling for long-running workloads
Scaling recruiting infrastructure
Scaling kubernetes clusters for thousands of resources
Adoption of ai in science at scale
Maintaining soc 2 compliance
Improving developer velocity
Finding top talent for ai and science roles
Regulatory compliance of ai toxicology reports
Benchmarking ai agent performance

▲Active Projects

Custom resource definitions and operators for ai agent lifecycles
Build an ai scientist
Experimentation platform development
Toxicology evaluation and validation pipelines for ai agents
Benchmarking ai agent performance
Production-ready integrations for client science workflows
Ai scientist development
Integrated research environment
Design biological research benchmarks
Commercial contract support

Hiring Activity

Accelerating30 roles · 30 in 30d

Department

Engineering

Research

Sales

Data

Operations

Ops

Seniority

Mid

Senior

Principal

Lead

Staff

Notable leadership hires: Head of DMPK

Company intelligence

Find more companies like Edison Scientific by tech stack, pain points and active projects

Get started free

About Edison Scientific

Edison Scientific is a scientist-led platform for R&D teams in regulated life sciences. The product bridges scientific discovery and development through AI that learns from organizational and experimental context, helping teams accelerate timelines from initial discovery through FDA-relevant milestones. Founded in 2025 as a spin-out from FutureHouse, Edison operates from San Francisco with 51–200 employees. Active development focuses on AI agent lifecycle management, toxicology evaluation pipelines, experimentation platforms, and production integrations for client science workflows. The company is early but capital-backed, signaling rapid scaling ahead.

HeadquartersSan Francisco, California

Company Size51–200 employees

Founded2025

Hiring MarketsUnited States

Frequently Asked Questions

What tech stack does Edison Scientific use?

Python, PyTorch, JAX, and Megatron-LM for AI; Node.js, FastAPI, GraphQL, React, and Next.js for backend and frontend; PostgreSQL, MySQL, MongoDB for data; AWS, GCP, Azure for cloud; Kubernetes, Terraform, Pulumi for infrastructure; Datadog, Prometheus, Grafana for observability.

What is Edison Scientific working on?

Custom operators for AI agent lifecycles, toxicology evaluation pipelines, experimentation platforms, production-ready integrations for client workflows, and benchmarking frameworks for AI agent performance in scientific contexts.

How this profile is built

Edison Scientific's technology stack, projects, and hiring signals are inferred from public hiring and company data — career pages, public listings, and company web presence — then clustered and de-duplicated. Figures are estimates that refresh over time. Read our full methodology →

This is not an official vendor or customer list. It is a technology-adoption signal inferred from public data, intended for B2B research.