Senior Full-Stack, Data and AI Agent Engineer

EPAM Systems Inc; United States; On-site; USD 150,000 - 210,000; Full time

Job summary

EPAM Systems is seeking a Senior Full-Stack, Data and AI Agent Engineer to join our team building a production agentic AI platform for upstream oil & gas operations. You will turn engineers' questions into safe, schema-aware SQL and multi-step reasoning workflows, while designing FastAPI endpoints and real-time SSE streams.

Key work includes integrating with AWS Bedrock, LangGraph, and MLflow, plus ensuring per-tenant isolation and robust observability.

Qualifications

• 3+ years backend engineering in Python with async/await and FastAPI or equivalent.
• Experience with AWS Bedrock, RDS, S3; no hardcoded credentials.
• Hands-on with LangGraph, LangChain, or similar agent orchestration.
• Strong SQL/PostgreSQL, including safety-critical code and tenant isolation.
• Familiarity with LLM APIs (Bedrock, OpenAI, Anthropic) and prompt engineering.
• CI/CD and test discipline: lint gates, pytest/Jest, GitHub Actions.
• Security-minded: tenant isolation, SELECT-only SQL, no cross-tenant leakage.
• Observability/eval mindset with MLflow or equivalent; monitor latency and tokens.
• Willingness to learn oil & gas terminology; client-facing English; Spanish nice-to-have.
• Nice-to-have TS/React/Next.js, SSE formats, payload shapes.
• Familiarity with uv, Ruff, and pytest.

Responsibilities

• Own the deepagents/LangGraph agent runtime, building and extending agent tools, subagents, and safety middleware.
• Design FastAPI endpoints with async I/O and SSE streaming for real-time agent responses.
• Write SELECT-only, injection-safe SQL generation against per-tenant PostgreSQL schemas.
• Integrate with AWS Bedrock and manage LLM factory/model routing.
• Instrument agent runs with MLflow and contribute to the eval framework.
• Work with pgvector, FAISS, and S3-backed document storage for RAG retrieval.

Skills

Python backend

Async/await

FastAPI

PostgreSQL

LangGraph/LangChain

AWS Bedrock

LLM APIs

CI/CD

Security & isolation

MLflow

pgvector / FAISS

Redis/APScheduler

TypeScript/React (nice-to-have)

Tools

psycopg3

pytest

Jest

LangChain

GitHub Actions

SSE streaming

pgvector

FAISS

S3

LangGraph

Job description

We are looking for a Senior Full-Stack, Data and AI Agent Engineer to join our team developing a production agentic AI platform for upstream oil & gas operations (ESP, gas lift, plunger lift, chemical injection). You will build and maintain the deep-agent core that turns petroleum engineers' natural-language questions into safe, schema-aware SQL and multi-step reasoning workflows.

Responsibilities

• Own the deepagents/LangGraph agent runtime, building and extending agent tools, subagents, and safety middleware
• Design FastAPI endpoints with async I/O (asyncpg/psycopg3) and SSE streaming for real-time agent responses
• Write SELECT-only, injection-safe SQL generation against per-tenant PostgreSQL schemas
• Integrate with AWS Bedrock (Claude via langchain-aws) and manage LLM factory/model routing
• Instrument agent runs with MLflow and contribute to the eval framework
• Work with pgvector, FAISS, and S3-backed document storage for RAG retrieval

Requirements

• 3+ years of backend engineering experience in Python, with strong async/await skills and FastAPI or a comparable framework
• Familiarity with AWS core services such as Bedrock, RDS, and S3, with a strict no-hardcoded-credentials approach
• Hands-on experience with LangGraph, LangChain, or a similar agent-orchestration framework
• Solid SQL/PostgreSQL experience, with comfort owning safety-critical code such as SQL injection prevention and tenant isolation
• Familiarity with LLM APIs (Bedrock, OpenAI, Anthropic) and prompt engineering
• Competency in CI/CD and test discipline, including lint/format gates, pytest/Jest, and GitHub Actions
• Security and tenant-isolation mindset, covering tenant isolation, SELECT-only SQL, and no cross-tenant leakage
• Observability/eval mindset using MLflow or equivalent, with quality, latency, and token measurement before shipping
• Flexibility to learn oil & gas / production-ops terminology (ESP, gas lift, decline curves) even without prior background
• English proficiency for client-facing communication; Spanish is a nice-to-have
• Nice to have Awareness of TypeScript/React/Next.js, including understanding of SSE event formats, content types, and payload shapes for cross-stack debugging
• Familiarity with uv, Ruff, and pytest
• Knowledge of MLflow
• Familiarity with Redis/APScheduler
• Background in oil & gas domain knowledge