Generative AI Engineer

If you have a strong track record of designing, building, and operating GenAI systems (across both text and images) in production environments, wed love to talk.\nWhat Youll Do\nOwn problems end‑to‑end : Translate business goals into technical plans; design pragmatic solutions; deliver production systems with measurable impact.\nBuild production back ends : Design and implement APIs and microservices (REST/gRPC) for GenAI workloads; containerize and orchestrate services (Docker/Kubernetes/ECS/EKS).\nShip on AWS : Leverage AWS (e.g., Lambda, ECS/EKS, S3, DynamoDB/RDS, API Gateway, SQS/SNS, CloudWatch) plus AI services (e.g., Bedrock , SageMaker ) to train, host, and integrate models.\nWork across modalities : Deliver features for text (LLMs/RAG) and images (VLMs/CV) including retrieval, embeddings, fine‑tuning/adapters, and evaluation pipelines.\nMake it observable : Instrument logging, metrics, and traces (OpenTelemetry/CloudWatch/Datadog/etc.); build dashboards, SLOs/SLIs, and alerts; own performance, reliability, and cost.\nValidate and govern : Implement offline/online evaluations, A/B tests, guardrails/red‑teaming, data and model quality checks, and safety/compliance gates.\nAutomate the path to prod : Establish CI/CD (GitHub Actions/CodePipeline), infrastructure as code (Terraform/CloudFormation), automated tests, and rollouts (canary/blue‑green).\nCollaborate without handoffs : Partner with product, domain experts, and downstream teams; document architecture; support launches; close the loop with data‑driven iteration.\nWhat Youve Shipped (Signals We’ll Look For)\nAt least one year owning a production GenAI or ML system (not a side project), professional experience building back‑end or ML‑powered products.\nServices you built that are running in production with users/traffic, clear SLIs/SLOs, and release/incident history.\nEvidence of quality: eval frameworks, regression tests, canary strategies, monitoring dashboards, cost/perf optimizations you introduced.\n\nRequired Experience\nGenAI foundation : LLMs/VLMs, embeddings, RAG, prompt orchestration, adapters/fine‑tuning, tokenization, latency/cost trade‑offs, content safety/guardrails.\nBack‑end & systems : Strong design of microservices, APIs, event‑driven patterns; data modeling across SQL/NoSQL; familiarity with vector databases .\nAWS & cloud infra : IAM/KMS/secrets, networking, containers/orchestration, CI/CD, IaC; operating services in AWS with cost/performance ownership.\nObservability & reliability : Logging, metrics, traces; performance profiling; incident response; chaos and load testing; availability and scaling strategies.\nLanguages & tooling : Proficient in Python (plus one of TypeScript/Go/Java); PyTorch/TensorFlow; Docker/Kubernetes; git; testing frameworks.\n\nDisclaimer: Firstsource follows a fair, transparent, and merit-based hiring process. We never ask for money at any stage. Beware of fraudulent offers and always verify through our official channels or @firstsource.com email addresses.