LLM Platform Engineer

Benchstrength; San Francisco (CA); Hybrid; USD 140,000 - 210,000; Full time

Benefits offered by this job

Equity

Generous PTO

Home office stipend

Job summary

AllSpice is seeking a senior software engineer to own the AI tooling, evaluation systems, and data platforms used by our Generative AI team. You will impact internal engineers and product managers, with opportunities to extend into customer-facing experiences.

You’ll shape workflows from discovery to rollout, grow domain expertise, and collaborate with product engineering to contribute to core features while preserving focus on tooling and reliability.

Qualifications

• 4+ years of software engineering experience, with ownership of tooling or data platforms.
• Strong SQL skills and experience designing data pipelines and storage schemas.
• Proficiency in Python scripting and automation for data pipelines.
• Experience building web applications or customer-facing tools using Vue/React/TypeScript.
• Hands-on with LLM-enabled systems, evaluation pipelines, and observability.
• Excellent communication skills and a customer-first mindset.

Responsibilities

• Own the tooling and data platforms for AllSpice's Generative AI team.
• Build internal applications and self-service tools for engineers and product managers.
• Develop data pipelines and integrations connecting internal systems, LLMs, and MCP tools.
• Maintain scripts and ensure data quality, schema evolution, and privacy.
• Document findings and communicate with technical and non-technical audiences.
• Champion code quality, testing, and reliability across the tooling stack.

Skills

SQL

Python

TypeScript

Vue/React

Data pipelines

Observability

Education

Bachelor's degree in a technology-related field

Tools

GitHub Actions

Docker

AWS

Job description

Help define the future of hardware development! At AllSpice, we're building the automation engine that powers the next generation of circuit design, enabling breakthroughs in smart vehicles, robotics, rockets, IoT, medical devices, and more.

We're creating the agile development environment for hardware engineers with a Git-friendly translation layer and automated CI/CD framework for native circuit designs. Think GitHub/GitLab + Copilot for electronics.

Our mission is to bring the same modern workflows that revolutionized software: version control, automation, and collaboration to the hardware world.

Learn more about our journey in TechCrunch and our recent Series A announcement.

If you're passionate about building tools that empower teams and delight customers, and you want to shape the foundation of modern hardware DevOps, we'd love to meet you!

Read more at https://allspice.io

What will you do?

This is a senior, high-impact software engineering role at the intersection of AI platform development, data engineering, and full-stack product development. You will own the tooling, evaluation systems, and data platforms that make AllSpice's Generative AI team more effective. Your primary users will be internal engineers and product managers, with opportunities to build customer-facing experiences where they support our AI products.

We're looking for someone who can identify the right problems, make thoughtful technical decisions, and take solutions from discovery through implementation, adoption, and ongoing improvement. As you build domain knowledge and establish the AI platform foundation, we'd also like you to grow into contributing directly to AllSpice’s core product alongside the product engineering team. This is an aspirational growth opportunity, not a prerequisite or the role’s initial primary focus.

You will:

• Partner with engineers and product managers to discover workflow bottlenecks, prioritize the highest-impact problems, and own solutions end-to-end - from requirements and architecture to rollout and iteration.

• Own the evaluation and observability layer for our AI products, including automated and human review workflows, regression detection, and comparisons across models, prompts, and pipelines. Make quality, cost, latency, usage, and customer outcomes visible and actionable.

• Build usable internal applications and self-service tools that help engineers and product managers inspect AI results, investigate failures, and provide feedback. Where useful, extend these capabilities into customer-facing experiences.

• Build and maintain scripts, reliable data pipelines, and integrations that connect internal systems, third-party services, LLMs, and Model Context Protocol (MCP) tools, with attention to data quality, schema evolution, and customer data privacy.

• Write clear documentation and communicate findings to both technical and non-technical audiences, including leadership and customers.

• Champion code quality, testing, and reliability across the tooling stack through reviews, mentorship, and thoughtful design.

Example Projects

• Build an internal evaluation workspace for DRCY, our AI design review product, that captures run configurations, inputs, outputs, and telemetry; compares model and prompt changes; and tracks quality, false positives, cost, and latency over time.

• Develop review workflows that let engineers and product managers inspect AI-generated findings, apply scoring rubrics, override automated scores, and investigate regressions—with opportunities to surface relevant insights and feedback workflows to customers.

• Create and maintain data pipelines that aggregate evaluation results, customer feedback, and telemetry into a unified analytics layer.

• Build scripting tools and automations to streamline the Gen AI team's development and testing workflows (e.g., batch evaluation runs, report generation).

• Design and implement RAG-based information repositories and tooling to capture component data and domain knowledge for use by our suite of AI products.

• Integrate with CI/CD systems (GitHub Actions) to surface AI insights directly in engineering workflows.

• Partner with infrastructure to measure cache effectiveness, usage, and cost, and use those insights to improve the reliability and efficiency of AI workflows.

• As an aspirational next step, contribute directly to core AllSpice product features—for example, integrating AI review findings and feedback into existing engineering workflows—in partnership with the product engineering team.

Our stack

• Languages: Python, TypeScript/JavaScript, SQL

• Analytics & Data: SQL databases (Postgres), Jupyter Notebooks, business intelligence/analytics tools

• AI/LLM: Langchain, Pydantic AI, Langfuse

• Web: Vue/TypeScript front-end, common web architectures and APIs

• Infrastructure: GitHub Actions for CI/CD, Docker Swarm & Terraform, AWS

• Our Hub application is a soft fork of Gitea

  • Go [server-side]

  • Vue/TypeScript Front-end

Expectations

Our ideal candidate should have the following:

• 4+ years of software engineering experience with a focus on tooling, data platforms, or full-stack web development, with demonstrated senior-level ownership of systems or products that other people depend on.

• Strong SQL skills and experience designing data pipelines and storage schemas, handling data quality and schema evolution, and using data/analytics tools (e.g., Jupyter, Metabase, Looker, or similar BI platforms).

• Proficiency in Python scripting and automation for data pipelines, integrations, and internal tools.

• Experience building web applications or customer-facing tools using modern frameworks (e.g., Vue, React, TypeScript).

• Practical experience shipping or maintaining LLM-enabled systems and building infrastructure around them, such as evaluation pipelines, observability, or integrations. You understand non-determinism, quality and reliability challenges, and tradeoffs across cost, latency, and model performance.

• Excellent communication skills - you can translate technical findings into clear insights for engineers, product managers, and customers alike.

• A customer-first mindset: you think about the end-user experience and proactively seek feedback to improve the tools you build.

• Bachelor's degree or higher in a technology-related field, or equivalent practical experience.

• Project management skills with the ability to lead and collaborate with varied stakeholders to move projects forward.

• Comfort with ambiguity and autonomy: you independently identify problems, propose solutions, make pragmatic build-vs.-buy and architectural decisions, and bring stakeholders along in a fast-moving environment.

• (Preferred) Availability to work out of our flex offices in San Francisco or Boston on a hybrid basis (2-3 times/week), but we are flexible for the right candidate.

Relevant Skills

You don't necessarily have to be proficient in all of these, but the more you know, the better:

• Python, TypeScript, SQL

• Business analytics and data visualization tools

• Langchain, Pydantic AI, and LLM evaluation frameworks

• GitHub Actions development and CI/CD

• Common web architectures, APIs, and front-end frameworks (Vue, React)

• Monitoring and observability tools (e.g., Langfuse, DataDog, Grafana)

• Project management and stakeholder communication

Benefits

• Supportive and smart colleagues, flexible work, opportunity to make a big impact, competitive salary & equity, health, dental, vision, generous PTO, home office stipend.