Sr Engineer Python AI

">Sr Engineer Python AI

5-8 Years Navi Mumbai

• Python
• RAG
• MCP
• LLM/GenAI
• SQL/PostgreSQL
• FastAPI

Senior Backend Engineer - Python/FastAPI AI/Agentic Systems

Skills Required

Backend Core (Python)

• Proficiency in Python 3.13, async/await - async-first codebase (SQLAlchemy w/ asyncpg/psycopg[binary]); minimal tolerance for sync code

• Advanced knowledge of FastAPI - routes, dependency injection, Pydantic v2 schemas, exception handlers

• SQLAlchemy 2.0 (async ORM) - rich domain models, business logic on the model layer (not repository/service wrappers)

• Advanced knowledge of API security, authentication and authorization

• Experience developing RESTful web APIs and real-time WebSocket services

• Familiarity with background job, queue, and task scheduling frameworks: APScheduler and AWS SQS (boto3)

• Advanced knowledge of data structures and algorithms with attention to time/space complexity

• PostgreSQL - schema design and hand-written idempotent SQL migrations (no ORM migration tooling)

• uv for dependency management, ruff for lint/format, pytest (async mode, AWS mocking)

• Comfort working within a strongly convention-driven codebase (documented architectural standards, not tribal knowledge)

AI / LLM Integration

• Agent framework experience - building agents, tool-calling loops, model configuration

• Model Context Protocol (MCP) - hands-on experience building or exposing MCP tools/servers, including how tool metadata (names/descriptions) shapes what an LLM sees

• AWS Bedrock - model invocation, plus familiarity with agent runtime deployment patterns (containerized, separate from core API/worker infra)

• Prompt engineering lifecycle management - designing prompts that work consistently across multiple delivery mechanisms (template-based vs. managed prompt services)

• RAG technique fluency - chunking/embedding strategy, vector store integration, retrieval-augmented prompt construction

• Guardrails - experience with LLM safety/guardrail configuration in production paths

• Queue/async job design - decoupled worker patterns (SQS/Redis-class tooling); understanding job dispatch, retries, DLQ semantics

• Experience integrating LLMs via the OpenAI SDK and prompt orchestration, tools/agents, and token budgeting

Database Caching

• Advanced knowledge of SQL and PostgreSQL

• Experience with database performance tuning, query optimization, and connection-pool sizing

• Proficiency with Redis for caching, session/state storage, and ephemeral coordination

Distributed Systems Architecture

• Understanding of fundamental design principles behind scalable, high-performing applications and how they fit into a larger microservices/distributed system

• Understanding of multi-tenant architecture - tenant isolation, data segregation, and shared-infrastructure patterns

• Understanding of asynchronous and event-driven processing, publish/subscribe, at-least-once delivery, dead-letter handling, idempotency, and eventual consistency

• Familiarity with resilience patterns - circuit breakers, retries with back-off, and graceful degradation across service boundaries

Platform Conventions to Onboard

• Queue-driven architecture (API enqueues, worker dispatches - no direct execution)

• Config-driven system design (DB-backed operational config with fallback defaults)

• Working within a shared internal scaffolding/template system and layered documentation set

Infrastructure DevOps

• Experience with CI/CD via Bitbucket Pipelines and artifact management

• Familiarity with AWS services: S3, SQS, SES, STS and similar

• Familiarity with monitoring/observability tooling (Splunk experience a plus)

• Familiarity with security scans, penetration tests, and vulnerability remediation

Collaboration Tooling

• Familiarity with the Atlassian suite: Jira, Confluence, Bitbucket

• Experience with Git branching strategies, pull-request workflows, and code review practices

• Experience working in remote and async-first collaboration environments

Nice-to-Have / Seniority-Gated

• Auth internals (Cognito-based auth, JWT scope middleware) - only for engineers touching auth/registration flows

• AWS deployment familiarity (SQS/S3/ECS Fargate) - abstracted by internal tooling, so lower priority

• Experience with lightweight local agent-testing UIs (e.g., Chainlit) - low priority

Experience Requirements

• 5+ years of experience in software engineering

• 3+ years of experience in a SaaS organization

• 3+ years Python backend, ideally async FastAPI/SQLAlchemy specifically (not just Flask/Django sync experience)

• Hands-on LLM/agent-building experience (built and iterated on real tool-calling agents - framework itself matters less than depth)

• RAG implementation experience (vector store + retrieval) is a strong differentiator

• Proven experience building Python/FastAPI backend services

• Experience with event-driven architectures using message brokers and/or cloud queues (SQS)

• Experience integrating LLM/RAG capabilities into production applications is a strong plus

• Experience with SOC 2 or similar compliance frameworks

• Experience working remotely with distributed teams across timezones

Hiring signal: Candidates with real MCP experience (not just generic "function calling") and exposure to managed agent runtime environments (e.g., Bedrock AgentCore or equivalent) will ramp faster than generalist LLM-app engineers.

Tasks Responsibilities

Backend Services - Ownership

• Design, build, and maintain efficient, reusable, reliable, and scalable FastAPI services - REST + WebSocket APIs backed by PostgreSQL and Redis

• Own and evolve the APIs supporting web clients and the mobile layer - contract design, versioning, and backward compatibility

• Build and maintain background jobs, schedulers, batch/cron tasks, and event consumers/publishers across RabbitMQ and SQS

• Uphold multi-tenant architecture principles and tenant-scoped data access across all backend work

• Identify performance bottlenecks and develop solutions in a high-traffic multi-tenant SaaS environment

AI / Assistant Layer

• Build and maintain LLM-powered features - LangChain/OpenAI orchestration, MCP tool integrations, and RAG pipelines over pgvector/FAISS

• Own embedding, retrieval, and prompt strategies; monitor quality, latency, and token cost

• Build and iterate on real tool-calling agents using agent frameworks (Bedrock AgentCore or equivalent)

• Implement guardrails and LLM safety configuration in production paths

• Design and maintain queue/async job patterns for decoupled agent worker execution

Architecture Design

• Translate business requirements into backend and AI solutions spanning FastAPI services and agentic layers

• Create design documents from software requirements and maintain architecture documentation

• Apply scalability and resilience methodologies (caching, queuing, circuit breaking) in a consistent and robust manner

• Define and evolve backend engineering standards (API contracts, state boundaries, agent patterns) as the team scales

Security Compliance

• Own authentication/authorization integration end-to-end - Keycloak OIDC and custom JWT/cookie flows across services

• Identify and remediate common API vulnerabilities within defined SLAs

• Participate in quarterly SOC 2 compliance activities including static, dynamic, and 3rd-party library scans

• Apply security-first development practices and dependency hygiene across the stack

Incident Response

• Support incident response and troubleshooting for enterprise clients within defined SLAs

• Use structured logs for diagnosis and post-incident review

• Identify root causes and implement permanent fixes for production issues

Team Collaboration

• Collaborate closely with the US-based oversight engineer, the technical architect, and peer engineers

• Participate in code reviews and establish code-quality standards

• Mentor and guide mid-level engineers as the team scales

• Work effectively in a remote, async-first environment across timezones

Disclaimer: This job posting has been aggregated from external source. Role details, content, and availability are subject to change. Applicants are advised to confirm the latest information directly on the company website before applying.