Plato.so — Software Engineer

Type: Full-time | On-site | San Francisco, CA Compensation: $180K–$280K + competitive equity Hiring count: 2–3 Visa sponsorship: No — not available Reports to: Pranav Putta, Co-Founder / CTO

About Plato.so

Plato is an applied research lab building the foundational infrastructure to train specialized AI agents. It turns real-world data streams into high-fidelity simulated environments that generate the training signal needed to make capable models, supporting frontier labs, hyperscalers, and enterprises building AI systems for complex, high-stakes work. Compute and algorithms are commoditizing fast; RL data remains the bottleneck, and Plato is built to scale training environments automatically from proprietary real-world data.

Founded: 2025 | Team size: 10 | Total funding: Raising Series A (~$18M round closing) Industry: AI, Data, Devtools Website: plato.so Office: FiDi, San Francisco

Why Candidates Should Join

• You're building the training loop — not wrapping it: Plato's software engineers work at the core of the RL pipeline. Not dashboards on top of someone else's model — the systems that actually generate training signal for frontier AI agents.
• Customers include OpenAI and Amazon AGI Labs: Not a research project searching for product-market fit. Plato already ships software that the most important AI labs depend on daily.
• Massive scope and ownership at a 10-person team: Touch backend systems, data pipelines, automation infrastructure, internal tools, and customer-facing prototypes. Flat structure means reporting directly to founders with full autonomy to drive projects end-to-end.
• Founder-led technical culture: Work alongside researchers and infra engineers in a tight feedback loop — no layers of management or PMs between you and the problem.

Intake Call Summary

• Company: Builds simulation environments for training LLMs. Closing a ~$18M Series A. Main customers include OpenAI and Amazon AGI Labs.
• Role: Maintain infrastructure and automate post-training pipelines; scale and optimize systems for high concurrency.
• Candidate: Wants engineers who can scale systems, with Python and Rust experience. Back-end skills preferred; CS degree valued but not mandatory.
• Team: Flat structure; reports directly to a co-founder or other leader. 10 people, all technical, mix of infrastructure and research engineering.
• Comp/logistics: Salary $180K–$300K (intake range) with competitive equity. Fully onsite in SF; no visa sponsorship currently offered.
• Timeline: 2–3 hires over the next two months. Fast cycle (1–2 weeks): intro chat, technical round, 2–3 day work trial.
• Pain points: Finding candidates at the right experience level who are genuinely interested in the maintenance side and comfortable with high-stakes infrastructure management.
• Ideal profile: Dynamic individuals comfortable with AI projects and high-throughput experiments; experience converting real-world data streams into simulations is a plus.

Note: intake salary range is $180K–$300K; the structured role field lists $180K–$280K. Role Details below reflects the structured field.

The Role

A backend-leaning software engineer with AI-project experience who can scale and automate Plato's post-training pipeline in a fast-moving, research-adjacent environment — optimizing systems for scale, bringing down cost, and iterating quickly on Python-based stacks. Bonus for growth-stage infra scaling or multi-agent coordination experience.

What You'll Be Doing

• Build and optimize automation pipelines that streamline the post-training stack for scale and cost efficiency
• Maintain and support high-concurrency infrastructure powering customer training pipelines
• Work with researchers and co-founders to turn experimental workflows into robust, production-ready systems
• Develop backend services and APIs for environment generation, trace ingestion, and telemetry
• Collaborate on parallelization and coordination of multiple agents across distributed systems
• Ship pragmatic, high-quality software in a flat, deeply technical ~10-person team

Tech stack: Python, Rust, FastAPI, TypeScript, Agent SDKs, Redis, Distributed Systems, CPU-based Infrastructure

Qualifications

Seniority

• 0–8 years of experience in software engineering, backend, or AI-adjacent work [Required]

Work Experience

• Currently working on AI/agent-related projects [Must have]
• Experience scaling systems at startups or big tech [Required]
• Built backend systems, data pipelines, or automation infrastructure [Required]
• Worked with large codebases and iterated quickly in Python [Required]
• Comfortable navigating and contributing to large codebases [Required]

Education

• Computer Science degree preferred [Strongly preferred]

Hard Skills

• Strong Python proficiency for rapid iteration [Must have]
• Experience with high-concurrency backend infrastructure (FastAPI, queuing, Redis) [Required]
• Experience with agent SDKs, LLM tooling, or RL pipelines [Strongly preferred]

Traits to Avoid

• Primarily a manager; not hands-on with code daily

Role Details

Salary$180K–$280KEquityCompetitive equityOn-site policyOn-site in San Francisco. Relocation support available for the right candidate.Visa sponsorshipNot availableEmployment typeFull-timeLocationSan Francisco, CA

Screening Questions

1. Are you able to work on-site full-time in San Francisco? If you're not currently based there, are you willing to relocate?
2. Do you require visa sponsorship to work in the United States?
3. Can the candidate be on-site? If not, is the candidate willing to relocate?
4. What is their salary expectation?
5. How actively is this candidate exploring new opportunities?

Interview Process

Stage 1 — Submit candidate After submitting, you're notified if the hiring manager wants to proceed.

Stage 2 — Intro Chat (30 minutes) A 30-minute introductory conversation with a founding team member. Covers background, motivations, interest in Plato, and general fit, and confirms logistics like location/relocation willingness and visa requirements. Goal: cultural alignment, communication skills, and genuine interest in the problem space.

Stage 3 — Technical Interview (30–60 minutes) Conducted by Pranav or one of Plato's engineers. Format is an experimental design question — typically a real problem Plato has hit, such as a parallelization issue coordinating multiple agents. Candidate and interviewer work through it together, covering baseline experiments, ablations, and experimental design. Evaluates depth of understanding, independent thinking, engagement, and fluid intelligence. Not a traditional coding interview — a collaborative problem-solving exercise.

Stage 4 — In-Person Work Trial (1–3 days) On-site work trial at Plato's SF office with the engineering team, on real or representative tasks. Primary signal-gathering stage: technical ability, iteration speed, working style, team fit. Minimum one full day, with flexibility for candidates balancing other commitments.

Stage 5 — Offer Extended

Stage 6 — Candidate Hired Candidate accepts the offer and starts.

Ideal Companies & Backgrounds

Updated Jun 8, 2026

Frontier AI labs and agent-focused startups (LLM agents, RL, post-training, computer use) OpenAI, Deepmindz Innovations, xAI, Cognition, Adept AI, Sierra Air Conditioning Inc., Factory AI, Daytona, E2B

High-growth Series A–C startups with strong backend/platform engineering teams (fast-moving, engineers wear many hats and scale systems) Dagster Labs, Hexaware Technologies, Retool, Railwaymen, Render, Vercel, Supabase, NEON | An IPG Health Company, Temporal Technologies, Prefect

Big tech companies with AI/agent or distributed systems teams (1–3 years, AI-related projects at scale) Google DeepMind, NVIDIA, Apple, Meta

Likely Paraform name/logo mismatches in the lists above (auto-matched to the wrong entities): "Sierra Air Conditioning Inc." almost certainly means Sierra (the AI agent company); "Deepmindz Innovations," "Hexaware Technologies," "Railwaymen," and "NEON | An IPG Health Company" appear to be mis-resolved (e.g., Neon, the Postgres company). Sanity-check before sourcing against these names.

Ideal Candidate Profiles

For reference only — do not source these specific profiles.

Aaditya Yadav — LinkedIn Software Engineering at the University of Waterloo | Waterloo, Canada

• (No fit reasons provided by the hiring manager.)

Rejected Candidate Feedback

• Candidates must show clear examples of hands-on AI agent/LLM infrastructure work, with direct experience building and scaling high-concurrency backend systems (e.g. FastAPI, Redis, queuing).
• Source candidates who can demonstrate rapid iteration in large Python codebases at growth-stage startups or big tech — not just theoretical or academic backgrounds.
• Prioritize candidates with explicit, verifiable detail on distributed agent coordination and RL pipeline implementations over vague or unsupported claims.
• Fraud watch: A recent submission was rejected as "the profile appears fraudulent." Profiles on this role have been flagged for fabrication — apply extra verification scrutiny.