About Sira AI
Sira AI is building an AI Agent Performance Platform.
The product helps teams observe, test, compare, optimize, experiment, and fix AI agents across every
prompt, model, workflow, and tool change.
The initial MVP helps AI teams upload or ingest agent runs, inspect what happened inside each run, compare one agent version against another, detect regressions, analyze cost and latency, and generate a clear decision report showing what improved, what broke, and what needs to be fixed.
The goal is simple: help teams understand whether their AI agent is working, whether the next version is better, and what to optimize before it reaches users.
Por favor, lea detenidamente la información de esta oferta de empleo para entender exactamente qué se espera de los posibles candidatos.
Location & Hours
Remote. We are open to candidates based in India, Eastern Europe, Europe, Latin America, South America, or similar global talent markets.
This is a full-time contract role, Monday–Friday, approximately 9:00 AM–5:00 PM U.S. Eastern Time, or equivalent hours with at least 3–4 hours of overlap with U.S. Eastern Time each business day.
Role Details
-Contract to founding engineer.
-We are looking for a senior-level builder with roughly 5 or more years of software engineering experience.
-The ideal candidate has at least 1 to 2 years of practical experience building with LLMs, AI agents, RAG
systems, evals, observability, analytics dashboards, or AI SaaS products.
-Formal credentials matter less than proof of work, speed of execution, product judgment, and the ability to build clean software without a large team.Responsibilities
-You will help design and build the first working MVP for Sira from the ground up.
-You will build trace ingestion through a JSON upload flow and a basic API endpoint so users can send
agent run data into the platform.
-You will design the database schema for users, workspaces, projects, agents, versions, runs, trace steps,
tool calls, errors, metrics, evaluations, and reports.
-You will build a dashboard where users can view runs, inspect trace steps, filter by version, identify failed
cases, and understand what happened inside an agent execution.
-You will build version comparison logic that compares two groups of agent runs across success rate, cost,
latency, error rate, tool usage, regression cases, and improvement cases.
-You will build a simple release or performance report that explains whether a new version looks better,
worse, or risky compared with the previous version.
-You will integrate lightweight LLM analysis to summarize failures, cluster common issues, and suggest
likely fixes or next tests.
-You will set up a clean deployment flow, environment variables, documentation, and a handoff-ready
codebase that another engineer could understand and extend.
-You will work closely with the founder to turn user feedback from design partners into simple product
improvements without overbuilding.
Required Skills
-Strong full-stack engineering ability with experience building real applications, not just scripts or
notebooks.
-Strong backend experience with Python and FastAPI, or Node.js with Express or NestJS. FastAPI is
preferred.
-Experience with React, Next.js, TypeScript, and modern dashboard-style user interfaces.
-Experience with PostgreSQL and database schema design. Experience with SQLAlchemy, Prisma, Drizzle, or similar ORM tools is useful.
-Comfortable working with structured JSON event data, logs, traces, metrics, and analytics-style data
models.
-Experience building REST APIs, authentication flows, admin dashboards, and internal or customer-facing SaaS products.
-Practical experience using LLM APIs such as OpenAI, Anthropic, or similar providers.
-Familiarity with LangChain, LangGraph, OpenAI Agents SDK, CrewAI, AutoGen, RAG systems, prompt
evaluation, or AI-agent workflows is highly valuable.
-Comfort with Docker, GitHub, environment variables, deployment basics, and clean project
documentation.
-Ability to move fast, communicate clearly, make tradeoffs, and avoid unnecessary complexity.
Preferred Tools & Technologies
-Frontend: React, Next.js, TypeScript, Tailwind CSS, shadcn/ui, dashboard UI patterns.
-Backend: Python, FastAPI, Node.js, Express, NestJS.
-Database: PostgreSQL, Supabase, SQLAlchemy, Prisma, Drizzle, Alembic, migrations.
-AI and LLM tooling: OpenAI API, Anthropic, LangChain, LangGraph, RAG, evals, prompt testing, agent
traces, LLM observability.
-Infrastructure: Docker, GitHub, Vercel, Render, Railway, Supabase, AWS basics, CI basics.
-Data and analytics: JSON traces, event data, cost calculations, latency metrics, success rates, failure
analysis, reporting.
Compensation
Before being offered the contractor role to build the MVP, you will be assigned a paid technical work sample of $500–$1,000, depending on the scope and candidate experience.
If offered the role, you will be assigned to build the MVP in 1 month, with $5,000 USD fixed for the initial MVP build.
Validation Retainer: $500–$1,000/month for light bug fixes, small improvements, weekly support, etc. while the founder validates the product.
If Validated: Potential longer-term full-time founding engineer role at $5,500–$7,000/month USD + 1% vested equity. Equity is not granted upfront. Equity would be vested and tied to a longer-term commitment, contribution, and continued progress.
Hiring Process
We'll start by reviewing your resume, GitHub, portfolio, past projects, and relevant AI or SaaS experience. Strong candidates will be invited to a short call to discuss their background, product thinking, and approach to building the MVP.
Final candidates will complete a paid technical work sample based on a simplified version of Sira's core workflow: ingesting two sets of AI-agent runs, comparing version performance, identifying failures, and generating a basic performance report. After the work sample, we'll select one engineer for the 4-week MVP. sprint.
Please apply through this posting and include your resume/CV, LinkedIn, GitHub, portfolio, or relevant product links. xohynlm Shortlisted candidates will be contacted by email for next steps.