Senior Full Stack AI Product Engineer

Bain & Company•Austin, TX

4d•$140,875 - $169,250•Hybrid

About The Position

Bain & Company is developing a suite of cutting-edge data and software solutions designed to revolutionize how the private equity industry uses data for investment insights and decision-making. The PEG Innovation team's mission is to create analytical solutions for Bain clients, teams, and the broader institutional investor space using proprietary software and data products. This includes the development, commercialization, and daily management of Bain's proprietary datasets, data, and software businesses. Senior Full-Stack AI Product Engineers design, build, and operate end-to-end AI product experiences across the PE due diligence platform, hosted on Microsoft Azure. This role sits at the intersection of product engineering and applied AI, combining strong full-stack software engineering with deep practical knowledge of how LLMs and agentic systems behave in production.

Requirements

Bachelor’s degree in Computer Science, Engineering, Information Systems, Data Science, or a related field, or equivalent practical experience.
6+ years of experience building production software, including significant experience delivering full-stack applications and/or AI-enabled systems in production environments.
Demonstrated experience shipping user-facing AI product features end-to-end, from backend orchestration and service design through frontend implementation and rollout.
Experience delivering agentic systems into production, including tool calling, multi-step workflows, RAG, structured output handling, and measurable quality controls.
Experience designing and operating evaluation frameworks, including golden datasets, regression gates, qualitative evaluation approaches, and CI controls that prevent quality regressions.
Experience operating distributed services in containerized environments such as Docker and Kubernetes, with ownership of monitoring, reliability, and incident response.
Demonstrated ability to mentor engineers and raise engineering standards through code review, design guidance, shared conventions, and strong technical judgement.
Strong experience building modern full-stack applications with robust frontend architecture and backend integration.
Strong frontend engineering skills, including React and/or Next.js, TypeScript, component-based UI development, API integration, and application state management.
Experience building polished product experiences for complex workflows, including tables, document-centric interfaces, review flows, workspace patterns, and real-time or streaming interactions.
Ability to design UX for AI systems, including confidence indicators, citations/source grounding, fallback states, edit/retry patterns, and human review steps.
Strong product sense and judgement in translating non-deterministic AI behavior into usable and trustworthy product experiences.
Strong Python proficiency, including FastAPI, Pydantic v2, async patterns, pytest, Ruff, and mypy (strict).
Deep hands-on experience with LangChain and/or LangGraph, including stateful graph construction, tool integration, checkpointing, HITL interrupts, and streaming patterns.
Familiarity with Google ADK or equivalent agentic orchestration frameworks.
Experience with Temporal, including workflow/activity authoring, signal/query handling, safe versioning strategies, and durable execution patterns.
Strong prompt engineering skills, including structured output design, system prompt construction, instruction clarity, sampling configuration, and multi-turn context management.
Experience designing and implementing RAG pipelines, including chunking, embedding selection, vector store integration, re-ranking, and retrieval quality evaluation.
Experience designing LLM evaluation systems, including golden dataset design, metric definition, LangSmith/Braintrust integration, and regression gate implementation.
Experience with context window management, including token budgeting, sliding-window truncation, semantic compression, and tool-call state persistence.
Experience with vector databases such as pgvector and/or Azure AI Search, including schema design, recall/precision tuning, and embedding lifecycle management.
Experience containerizing and operating AI workloads using Docker and Kubernetes, including Helm charts and autoscaling concepts.
Uses AI coding assistants such as Cursor and GitHub Copilot as a core part of the development workflow, while applying strong judgement about where generated code is reliable versus where it requires scrutiny.
Designs multi-agent systems with clear orchestration logic, specialist sub-agent boundaries, tool interfaces, and failure-handling patterns.
Builds and maintains evaluation pipelines that combine deterministic metrics with LLM-as-judge patterns for qualitative assessment.
Capable of evaluating AI-generated infrastructure and workflow code, including Kubernetes manifests, Terraform, prompts, and agent graphs, for correctness, safety, and operability before production release.
Treats non-determinism as a first-class engineering challenge and designs systems that degrade gracefully when model outputs are unexpected.
Writes evaluation tests before shipping new AI capabilities, not after.
Prototypes quickly using AI tooling, but validates every production artifact against defined quality gates before promotion.
Documents behavior specifications, tool contracts, and user-facing interaction patterns rather than leaving critical behavior implicit in code.

Responsibilities

Design and build end-to-end AI product features across backend services, orchestration layers, and frontend user experiences.
Develop analyst-facing and internal AI interfaces for workflows such as deal screening, commercial due diligence research, document extraction, and portfolio monitoring.
Build responsive, high-quality frontend experiences for streaming AI responses, structured outputs, source grounding, review and approval flows, and human-in-the-loop interactions.
Implement full-stack application patterns for chat, copilot, workspace, and review-based AI experiences, including state management, real-time updates, and error handling.
Partner closely with Product, Design, and domain stakeholders to translate ambiguous AI capabilities into intuitive, polished user experiences.
Design and maintain stable contracts between frontend applications and AI/backend services, ensuring outputs are structured, testable, and resilient.
Build contribution workflows and product surfaces for the Prompt Execution Sandbox and AI Artifact Studio, enabling safe and scalable use by non-engineers where required.
Ensure AI product features are accessible, observable, and production-ready, with strong attention to usability, reliability, and edge-case handling.
Design and build the Agent Gateway service, including inbound APIs, model routing (including Azure OpenAI Service endpoints), context management, response validation, and cost/audit logging.
Build and maintain LangGraph agent workflows for PE use cases, including streaming, tool-calling, multi-step execution, and human-in-the-loop interrupt patterns.
Integrate Temporal durable execution with LangGraph, including workflow and activity authoring, checkpointing strategies, retry and backoff policies, signal/query handling, and safe versioning for in-flight workflows.
Build and maintain AI platform services such as Agent Session Manager, Memory Service, HITL Coordination Service, and Feedback/Correction Service with production SLO expectations.
Design and implement RAG pipelines end-to-end, including chunking strategies, embedding model selection, vector store integration (Azure AI Search and/or pgvector), re-ranking, and retrieval quality evaluation.
Own evaluation and regression gates, including golden dataset management, metric definition, qualitative and quantitative evaluation, LangSmith/Braintrust integration, and CI enforcement on quality regressions.
Implement context window management strategies such as token budgeting, truncation/compression, and tool-call state persistence to support reliability in longer-running workflows.
Instrument AI services with structured logging, traces, and metrics, and define operational dashboards and alerts for latency, quality, cost, and failure signals.
Deploy and operate AI workloads in Kubernetes, including containerization, Helm-based deployment patterns, and autoscaling considerations where appropriate.
Mentor engineers and help set standards for production AI product engineering across testing, evaluation, documentation, safety constraints, and maintainability.
Partner with Data Platform on feature store access patterns, inference integration, schemas, data contracts, and SLAs.
Collaborate with Product Engineering and Design on AI feature surfacing, including streaming experiences, structured output rendering, citation and evidence UX, and HITL review interfaces.
Use AI coding assistants to accelerate prototyping and development, while validating all production artifacts against testing and evaluation gates before promotion.
Document agent behavior specifications, tool contracts, and product interaction patterns so behavior is explicit, reviewable, and maintainable.