Time since we first saw it
Day 41
Up longer than most. Similar jobs usually come down between day 1 and day 15.
From 1,368 similar postings that came down. Coming down is not always a hire.
Bengaluru, India
Live, first seen by us 41 days agoEmployment type in the employer’s posting: Full timeFrom the employer’s posting
3–6 years of relevant, hands-on experience building and shipping full-stack production applications — backend-heavy is fine, but you should be able to own a React feature without hand-holding. Bonus if some of that experience was at an early-stage product company.
These dates show when we found the posting. Check the employer’s website to confirm it is still accepting applications.
Day 41
Up longer than most. Similar jobs usually come down between day 1 and day 15.
From 1,368 similar postings that came down. Coming down is not always a hire.
1 open job
Open jobs by the week we first saw them, last 12 weeks.
Hiring recordAdd your resume and we’ll mark which of these tools you already use, and point out anything in the posting that could rule you out.
• Agentic AI: you’ve built agents that do more than answer questions — they plan, decide, call tools, and recover from failures. You understand the difference between single-step tool-calling and multi-step agent graphs (LangGraph, CrewAI, or similar), know how to design reliable tool interfaces, handle partial failures and retries gracefully, manage context window constraints across long-running tasks, and evaluate whether an agent actually completed its goal — not just whether it returned a response. Prompt engineering discipline (system prompts, few-shot examples, structured outputs) is assumed. • Frontend fundamental: you understand how the web actually works — HTML structure and semantics, CSS layout and specificity, JavaScript (ES6+) including the event loop, promises, and async/await. You know the full HTTP request/response cycle: methods, status codes, headers, cookies, CORS, and how a browser request travels from the user to your FastAPI service and back. You understand REST vs WebSocket vs SSE and can pick the right protocol for the job. You can read browser DevTools network traces, debug a broken API call from the frontend, and reason about where a problem lives without needing someone to hand you a reproduction. React or another framework is expected on top of this — but the fundamentals come first.
LinkedIn: Alethic Insights | LinkedIn If you love owning full-stack products end-to-end, thrive on building production-grade Python services and polished React interfaces, and get genuinely excited about LLMs and agents actually doing things — this role was written for you. At our end, we are building our flagship platform that blends data, workflow automation, and agentic AI to help enterprises make smarter decisions, faster. We are looking for a hands-on engineer with 3–6 years of relevant experience who can own services end-to-end: design, build, deploy, monitor, and iterate. You’ll ship production code, shape the architecture, and help us push the boundaries of what agentic AI can do inside a real product.
• Real ownership. You own modules end-to-end — design, ship, monitor, iterate. • True full-stack ownership. React frontends, Python APIs, cloud infra, AI pipelines, auth, observability — the whole stack, your call. • AI that’s more than hype. Build and own RAG pipelines, vector search, LLM integrations, and autonomous agents that do actual work inside a real enterprise product.
• Build and ship React interfaces that enterprise users actually love. Design and implement component hierarchies, manage state across complex workflows, integrate with REST APIs, handle async data gracefully, and make sure the UI stays responsive even when an AI agent is doing heavy work in the background. • Design and ship Python/FastAPI services that are built to last. Clean router structure, Pydantic-validated contracts, dependency-injected auth, background task queues (Celery/RQ), async I/O where it matters — and the discipline to write tests before you call something done. • Implement and own end-to-end RAG pipelines. Not just “plug in LangChain and call it done” — you’ll choose the right chunking strategy for the document type, select and fine-tune embedding models, stand up and query vector stores (Pinecone, pgvector, or Weaviate), add reranking, and instrument retrieval quality metrics so you know when the pipeline degrades in production.
• SQL & data modelling: PostgreSQL is your default. You write joins, window functions, CTEs, and subqueries without hesitation, you understand EXPLAIN ANALYZE output, you design schemas with multi-tenancy in mind, and you don’t wait for a DBA to tell you your query is slow. • Python & FastAPI: async/await, Pydantic v2, dependency injection, background tasks, structured logging, and the instinct to write a test alongside the feature — not a week later. You can explain the difference between asyncio concurrency and true parallelism, and you know when Celery is the right answer and when it’s overkill. • Auth & security fundamentals: JWT, OAuth 2.0, OIDC, RBAC, and the habit of auditing every endpoint for tenant isolation before merging. You’ve thought about prompt injection, you know what OWASP Top 10 means in a Python/React context, and you don’t store secrets in environment variables committed to git.
• Python & FastAPI: async/await, Pydantic v2, dependency injection, background tasks, structured logging, and the instinct to write a test alongside the feature — not a week later. You can explain the difference between asyncio concurrency and true parallelism, and you know when Celery is the right answer and when it’s overkill. • Auth & security fundamentals: JWT, OAuth 2.0, OIDC, RBAC, and the habit of auditing every endpoint for tenant isolation before merging. You’ve thought about prompt injection, you know what OWASP Top 10 means in a Python/React context, and you don’t store secrets in environment variables committed to git. • Systems thinking: you can sketch a multi-service architecture, reason about where state should live, identify single points of failure, and make deliberate tradeoffs between consistency and availability. When to use a queue, when to go synchronous, when Redis is the answer and when it’s a crutch — you have opinions and can defend them.
• Generative AI & RAG: you’ve built a RAG pipeline in production — not just followed a tutorial. You understand the tradeoffs between chunking strategies (fixed-size, semantic, document-structure-aware), have selected and compared embedding models, have stood up at least one vector store (Pinecone, pgvector, Weaviate, or Qdrant), and you know that retrieval quality requires measurement, not hope. Fluency with LLM APIs (OpenAI, Anthropic, AWS Bedrock) and at least one agent framework (LangGraph, LangChain, LlamaIndex) is expected. • SQL & data modelling: PostgreSQL is your default. You write joins, window functions, CTEs, and subqueries without hesitation, you understand EXPLAIN ANALYZE output, you design schemas with multi-tenancy in mind, and you don’t wait for a DBA to tell you your query is slow. • Python & FastAPI: async/await, Pydantic v2, dependency injection, background tasks, structured logging, and the instinct to write a test alongside the feature — not a week later. You can explain the difference between asyncio concurrency and true parallelism, and you know when Celery is the right answer and when it’s overkill.
Good-to-Have • React & TypeScript: functional components, custom hooks, Zustand or Redux for state, React Router, typed API calls with axios or react-query, and clean handling of loading/error/empty states. TypeScript by default. You don’t need a senior frontend engineer reviewing your code for correctness — just for taste. • LLM evaluation and observability: experience with RAGAS, LangSmith, Phoenix/Arize, or similar tools for measuring retrieval quality, hallucination rates, and agent task completion — not just “it looked right in the demo.”
What You’ll Actually Do • Own features end-to-end across the full stack. Take a feature from design doc to production — React UI, FastAPI service, database schema, AWS deployment, and monitoring — without handing off to another team. You are the team. • Build and ship React interfaces that enterprise users actually love. Design and implement component hierarchies, manage state across complex workflows, integrate with REST APIs, handle async data gracefully, and make sure the UI stays responsive even when an AI agent is doing heavy work in the background.
• Own the data layer from schema to query plan. Model relational data in PostgreSQL with multi-tenancy and row-level security in mind, write queries a DBA would be proud of, reach for Redis or DynamoDB when the access pattern demands it, and know how to spot an N+1 before it becomes a production incident. • Deploy and operate your own services on AWS. Containerize with Docker, orchestrate on ECS or Lambda, store secrets in Secrets Manager, wire IAM with least privilege, set up RDS with proper connection pooling, and build CloudWatch dashboards and alarms so you know when something is wrong before a user files a ticket. • Implement auth you’d stake your reputation on. JWT issuance and validation, OAuth 2.0 / OIDC flows, role-based access control, secure session management, and the habit of asking “can this endpoint be called by the wrong tenant?” before every PR merge.
Must-Have Skills • Generative AI & RAG: you’ve built a RAG pipeline in production — not just followed a tutorial. You understand the tradeoffs between chunking strategies (fixed-size, semantic, document-structure-aware), have selected and compared embedding models, have stood up at least one vector store (Pinecone, pgvector, Weaviate, or Qdrant), and you know that retrieval quality requires measurement, not hope. Fluency with LLM APIs (OpenAI, Anthropic, AWS Bedrock) and at least one agent framework (LangGraph, LangChain, LlamaIndex) is expected. • SQL & data modelling: PostgreSQL is your default. You write joins, window functions, CTEs, and subqueries without hesitation, you understand EXPLAIN ANALYZE output, you design schemas with multi-tenancy in mind, and you don’t wait for a DBA to tell you your query is slow.
• LLM evaluation and observability: experience with RAGAS, LangSmith, Phoenix/Arize, or similar tools for measuring retrieval quality, hallucination rates, and agent task completion — not just “it looked right in the demo.” • Data pipeline experience: ingestion, transformation, and enrichment pipelines (Airflow, AWS Glue, or even just well-structured Celery workflows) — particularly for feeding structured and unstructured data into AI systems at scale. • Infrastructure as code: Terraform, AWS CDK, or Pulumi — the ability to version and review infrastructure changes the same way you review code. Not required, but the team will love you for it.
• Data pipeline experience: ingestion, transformation, and enrichment pipelines (Airflow, AWS Glue, or even just well-structured Celery workflows) — particularly for feeding structured and unstructured data into AI systems at scale. • Infrastructure as code: Terraform, AWS CDK, or Pulumi — the ability to version and review infrastructure changes the same way you review code. Not required, but the team will love you for it. • Enterprise SaaS patterns: multi-tenancy, row-level security, usage metering, audit logging, SSO/SAML integration, or experience building for compliance-conscious enterprise buyers. You’ve shipped something to a customer who cares about SOC 2 or GDPR.
• Build agentic AI workflows that actually complete tasks. Design multi-step agent graphs using LangGraph or similar, wire up tool-calling against internal APIs and external services, handle failure modes and retry logic, and evaluate agent performance — not just “did it finish” but “did it finish correctly.” • Own the data layer from schema to query plan. Model relational data in PostgreSQL with multi-tenancy and row-level security in mind, write queries a DBA would be proud of, reach for Redis or DynamoDB when the access pattern demands it, and know how to spot an N+1 before it becomes a production incident. • Deploy and operate your own services on AWS. Containerize with Docker, orchestrate on ECS or Lambda, store secrets in Secrets Manager, wire IAM with least privilege, set up RDS with proper connection pooling, and build CloudWatch dashboards and alarms so you know when something is wrong before a user files a ticket.
• Enterprise SaaS patterns: multi-tenancy, row-level security, usage metering, audit logging, SSO/SAML integration, or experience building for compliance-conscious enterprise buyers. You’ve shipped something to a customer who cares about SOC 2 or GDPR. • NoSQL and search: DynamoDB for high-throughput key-value access patterns, Redis beyond basic caching (pub/sub, streams, sorted sets), or Elasticsearch/OpenSearch for full-text and hybrid search — knowing when to reach past PostgreSQL and why. • Prior startup or product company experience: you know what it means to own a feature end-to-end with limited resources, make pragmatic tradeoffs under time pressure, and ship to real users who notice when things break.
• Auth & security fundamentals: JWT, OAuth 2.0, OIDC, RBAC, and the habit of auditing every endpoint for tenant isolation before merging. You’ve thought about prompt injection, you know what OWASP Top 10 means in a Python/React context, and you don’t store secrets in environment variables committed to git. • Systems thinking: you can sketch a multi-service architecture, reason about where state should live, identify single points of failure, and make deliberate tradeoffs between consistency and availability. When to use a queue, when to go synchronous, when Redis is the answer and when it’s a crutch — you have opinions and can defend them. • Agentic AI: you’ve built agents that do more than answer questions — they plan, decide, call tools, and recover from failures. You understand the difference between single-step tool-calling and multi-step agent graphs (LangGraph, CrewAI, or similar), know how to design reliable tool interfaces, handle partial failures and retries gracefully, manage context window constraints across long-running tasks, and evaluate whether an agent actually completed its goal — not just whether it returned a response. Prompt engineering discipline (system prompts, few-shot examples, structured outputs) is assumed.
• Frontend fundamental: you understand how the web actually works — HTML structure and semantics, CSS layout and specificity, JavaScript (ES6+) including the event loop, promises, and async/await. You know the full HTTP request/response cycle: methods, status codes, headers, cookies, CORS, and how a browser request travels from the user to your FastAPI service and back. You understand REST vs WebSocket vs SSE and can pick the right protocol for the job. You can read browser DevTools network traces, debug a broken API call from the frontend, and reason about where a problem lives without needing someone to hand you a reproduction. React or another framework is expected on top of this — but the fundamentals come first. Good-to-Have
Complete your application on healthminds.zohorecruit.in. The employer’s form will show what is required.
Already applied? Track this application
New AI Engineer postings at the employers we follow are up 340% over the last 6 weeks compared with the 6 before.
Each bar is one week of new postings. The darkest bars are the last 6 weeks.
Counted from the hiring pages we’ve followed for at least five weeks, so employers we add don’t look like growth. Updated weekly.
We file this job under AI Engineer, by the work the posting describes.
AI engineers build products on top of AI models that already exist, most often large language models: they connect a model to a company's data and tools and ship the assistants, agents and search features people use.
Postings ask for strong Python or TypeScript, experience with model APIs, retrieval and agent frameworks, and ways to test whether the model's answers are good enough.
An ML engineer trains and serves models; an AI engineer usually starts from a model someone else trained and builds the product around it, though titles often mix the two.
Selected passages from the saved posting. Check the full description for conditions and exceptions.
No pay amount identified in the saved description.
Bengaluru, India
• Enterprise SaaS patterns: multi-tenancy, row-level security, usage metering, audit logging, SSO/SAML integration, or experience building for compliance-conscious enterprise buyers. You’ve shipped something to a customer who cares about SOC 2 or GDPR. • NoSQL and search: DynamoDB for high-throughput key-value access patterns, Redis beyond basic caching (pub/sub, streams, sorted sets), or Elasticsearch/OpenSearch for full-text and hybrid search — knowing when to reach past PostgreSQL and why. • Prior startup or product company experience: you know what it means to own a feature end-to-end with limited resources, make pragmatic tradeoffs under time pressure, and ship to real users who notice when things break.
No clear work-authorization passage found. Eligibility is unconfirmed.