Full-Stack AI Engineer
Location: Hybrid | Global | Flexible Schedule
Type: Contract
About Bridge BSS
Bridge BSS delivers cutting-edge software solutions, AI innovations, and IT staff augmentation for global enterprises.
We partner with renowned brands and foster an inclusive, high-energy, learning-driven culture.
The Role
We’re hiring a Full-Stack AI Engineer to build AI-powered products end-to-end.
You will actively use both Node.js (Fastify) and Python (FastAPI) to design, ship, and operate production services and modern web apps (Next.js/React + TypeScript).
You’ll collaborate with product, design, data science, and platform engineering to turn ambiguous problems into safe, delightful, and measurable user experiences.
What You’ll Do
Prompt Engineering & LLM Integration
- Design, test, and refine prompts (zero-shot, few-shot, chain-of-thought, ReAct) across models (GPT-4/4o, Claude, Mistral, Llama, etc.).
- Implement tool/function calling, structured outputs (typed/JSON schema), and guardrails for reliability, safety, and cost control.
- Build production RAG pipelines: chunking strategies, embeddings, retrieval orchestration, caching, and evaluation.
- Analyze outputs for hallucinations, bias, tone, factual accuracy, and business alignment.
Backend & Platform (Node + Python)
- Build and own polyglot services in Node.js/TypeScript (Fastify) and Python (FastAPI) with shared standards for auth, observability, and reliability.
- Implement OAuth2/JWT, rate limiting, retries/backoff, and real-time streaming via SSE/WebSockets across both runtimes.
- Integrate with vector databases (Pinecone, Weaviate, pgvector), relational (Postgres), non-relational (MongoDB), and caches.
- Ship with CI/CD, containerization (Docker), and cloud deploys (AWS/GCP/Azure); instrument cost/latency telemetry, logging, metrics, and tracing.
Frontend (Next.js)
- Build user-facing experiences with Next.js (TypeScript, App Router), React Server Components, and Server Actions.
- Deliver chat UIs, evaluators/annotators, admin dashboards, and prompt versioning UIs with real-time streaming.
- Ensure UX clarity, accessibility, and responsive performance; collaborate closely with design (shadcn/ui + Tailwind).
LLMOps, Quality, and Safety
- Maintain a version-controlled prompt library with documentation, test cases, and use-case mappings.
- Run structured experiments, A/B tests, and quantitative/qualitative evaluations; automate golden sets and regression checks.
- Apply responsible AI practices; mitigate misuse, prompt injection, toxic content, and data-privacy/PII risks.
- Stay current on emerging techniques and share best practices with the team.
Required Skills & Experience
Experience
- 3+ years professional full-stack development (shipping production APIs and web apps).
- 2+ years hands-on AI/LLM experience in production (prompt engineering, LLM APIs, RAG, evals, guardrails).
Languages & Frameworks
- Advanced JavaScript and TypeScript.
- Node.js with Fastify/Express (required).
- Python with FastAPI (required). You’ll use both Node and FastAPI regularly.
Backend & APIs
- REST and GraphQL; async I/O; OAuth2/JWT; rate limiting; retries/backoff.
- Real-time streaming with SSE and WebSockets; resilient job/queue patterns.
Data & ORMs
- Postgres and MongoDB; schema design and query optimization.
- ORM (required); experience with Prisma/TypeORM/Drizzle or SQLAlchemy is beneficial.
- Migrations (e.g., Prisma Migrate), connection pooling, and transaction patterns.
LLMs & Tooling
- OpenAI and Anthropic APIs; familiarity with Mistral and Llama ecosystems.
- Prompt strategies (zero/few-shot, CoT, ReAct), function calling, structured outputs, and guardrails.
- LangChain and/or LlamaIndex in production; prompt/versioning/eval tools (e.g., PromptLayer, Cursor, Lovely).
Retrieval & Evaluation
- Vector stores: Pinecone, Weaviate, pgvector.
- Embeddings, chunking, hybrid retrieval, caching; offline/online evaluation frameworks.
DevOps & Reliability
- Docker and Kubernetes; CI/CD pipelines.
- Observability: logging, metrics, tracing; cost and latency telemetry for LLM usage.
- Secrets management and secure configuration.
Security & Safety
- Content safety techniques, prompt-injection defenses.
- Data privacy and PII handling/compliance best practices.
Collaboration
- Git, code review, excellent written communication and documentation.
- Comfortable collaborating via Jira/Teams/Figma with cross-functional teams.
Why Bridge BSS
- Own high-impact, end-to-end AI product experiences.
- Collaborative, learning-driven culture with global teammates.
- Modern stack, real users, and meaningful scale.
.png)
