Senior Software Engineer — Full-Stack & Agentic AI

Overview

Senior Software Engineer — Full-Stack & Agentic AI

Steps AI · On-Site, Hyderabad, India · Full-Time

About Steps AI

Steps AI is an agentic AI platform that powers customer-facing AI agents for businesses across e-commerce, SaaS, healthcare, real estate, EdTech, and financial services. Our agents go live in under five minutes from a single URL, learn the business by ingesting websites, documents, FAQs, and product catalogs, and deploy across web, messaging, and social channels.

We are profitable, growing fast, and shipping production AI to customers every week.

The Role

We are hiring a Senior Software Engineer who can architect, build, and own product features across the full stack — from agentic AI to backend platform to customer-facing surfaces. This is a hands-on senior IC role with significant scope: you will work directly with the founders, set technical direction for major product areas, and ship features used by real customers within days of being designed.

The role spans four production surfaces:

• A conversational AI service that powers every customer interaction — agent runtime, tool ecosystem, retrieval, streaming, and human-in-the-loop flows
• A workflow and ingestion service handling long-running jobs — content acquisition, document understanding, embedding, and scheduled processing
• A core application backend — the system of record for tenants, integrations, billing, channels, and CRM-style entities
• A customer-facing layer — a dashboard for businesses to configure and operate their agents, plus an embeddable widget that runs inside customer websites

You will be expected to be productive across all four — not necessarily on day one, but within the first 30–60 days.

Resumes will only be reviewed if the application form is completed.

Application form: https://forms.gle/ciFTj8iS6j3TLVNt7

What You Will Own

Agentic AI Systems

• Agent runtime design — state, conditional routing, tool binding, message handling, escalation flows, and custom event emission
• Multi-provider LLM orchestration with per-persona model selection, fallback chains, and cost/latency tradeoffs
• A tool ecosystem spanning e-commerce platforms, CRM and helpdesk systems, productivity suites (mail, calendars, docs, knowledge bases), logistics carriers, web search, and channel-specific surfaces (rich messaging templates, human handoff). New providers should reach production in days, not weeks
• Hybrid retrieval — dense and sparse signals with cross-encoder reranking, recall/precision tuning, and ingestion-time quality controls
• Memory systems — long-term cross-session memory and short-term per-thread state, with clear contracts between them
• Streaming protocols — token-level delivery, intermediate event emission, interrupt and resume semantics, and structured terminal events

Workflow Orchestration & Data Engineering

• Long-running workflows for ingestion, crawling, Q&A and FAQ synthesis, product discovery, sentiment work, and outbound campaigns
• Worker topology — queue separation, concurrency controls, retry semantics, and end-to-end tracing
• Document understanding — multi-format parsing, OCR for scanned content, and layout-aware extraction at scale
• Adapter-based crawling with pluggable backends for varied site profiles
• Idempotent embedding and indexing pipelines

Backend Platform

• Backend modules covering agents, workspaces, RBAC, knowledge resources, channel integrations, CRM-style entities (leads, tickets, customers, live requests), analytics, billing, and webhooks
• Relational schema design at production scale
• OAuth 2.0 across many third-party providers, with encrypted token storage
• Messaging-platform integrations across major channels — including signed-state onboarding flows, subscription management, encrypted payloads, and webhook signature verification
• Subscription billing with multi-currency pricing, lifecycle management, and usage-based components
• Multi-tenant workspace RBAC with quota and feature guards

Frontend & Embeddable Widget

• A customer-facing dashboard covering agent building, knowledge management, integration marketplace, persona library, skills configuration, channels, CRM, unified inbox, real-time analytics, and billing
• Public agent surfaces with custom slugs and brand kits
• An embeddable widget delivered via CDN — host-page isolation, streaming, local caching, live-chat over a persistent connection, and a polished mobile experience
• Modern client- and server-state patterns with authenticated session management
• Data visualisation across the dashboard's analytics surfaces

Cloud, DevOps & Production

• Containerised local and staging environments
• Cloud deployment (AWS), CI/CD pipelines, multi-environment promotion (dev → staging → production)
• Operating relational, vector, cache, and object storage in production
• Error monitoring, distributed tracing, LLM-specific observability, and metrics
• Webhook signature verification, JWT and refresh-token rotation, and secrets management

Business → Code Translation

• Translate business requirements into shipped features without extensive hand-holding
• Architect new modules end-to-end (data model → API → workflow → AI tool → UI)
• Defend tradeoffs in writing through RFCs and design docs

Required Skills

You should have shipped these to production and be able to demonstrate them live:

Agent frameworks Production experience with LangGraph, LangChain, or comparable agent runtimes — tool calling, state machines, checkpointing, streaming, human-in-the-loop
LLM APIs OpenAI, Azure OpenAI, Anthropic, Google Gemini — direct integration and proxying patterns
Vector databases At least one of Milvus, Pinecone, Weaviate, or Qdrant — schema design, hybrid retrieval, idempotent upsert
Backend (Python) FastAPI, async I/O, Pydantic, structured logging
Backend (TypeScript) NestJS or comparable Node frameworks, an ORM (e.g. Prisma), validation libraries
Workflow orchestration Temporal, Airflow, Prefect, or equivalent at production scale
Database PostgreSQL — production-scale schema design, indexing, query optimisation
Frontend Modern React with a leading meta-framework (Next.js App Router or equivalent), a server-state library (TanStack Query or similar), a client-state library (Zustand, Jotai, or similar), Tailwind, and a headless component system (shadcn/ui or Radix)
Cloud AWS or GCP — deploy, monitor, scale, and debug in production
DevOps Docker, CI/CD, environment management, observability tooling

Strong Plus

• Messaging platform APIs at scale (WhatsApp Business, Instagram, Messenger) — including signed-state OAuth and webhook flows
• E-commerce platform APIs (Shopify Admin GraphQL with webhooks, WooCommerce REST, or similar)
• Subscription billing at scale (multi-currency, lifecycle, usage-based)
• Multi-tenant SaaS with workspace-level RBAC
• LLM observability — Langfuse, OpenTelemetry, and similar instrumentation
• Multimodal AI — vision, audio, voice — including transcription, vision-language models, and live multimodal streams
• Embeddable JavaScript widgets — host-page isolation, lightweight runtimes, CDN delivery
• Open-source contributions in the GenAI / agentic / web framework / orchestration ecosystem
• Published papers, patents, or talks in the GenAI space

Mindset We Hire For

• You ship rapidly. A well-defined feature goes from idea to production in days, not weeks.
• You read code faster than docs. When debugging a cross-service issue, you trace through the stack and find the root cause yourself.
• You design for failure. Every webhook is idempotent, every queue job is retried correctly, every external API has circuit-breakers, every prompt has guardrails.
• You measure what you ship. Token usage, latency, retrieval quality, and conversion rates are instrumented from day one.
• You take ownership. When something breaks, you fix it, write a postmortem, and prevent recurrence.
• You translate fuzzy business requirements into hard technical specs without multiple rounds of clarification.

Qualifications

• B.Tech / M.Tech / MS in CS, Software Engineering, AI/ML, or a related field
• 3+ years of professional engineering experience with a strong production track record
• Demonstrable experience shipping agentic AI systems and full-stack applications to real users
• On-site availability in Hyderabad — every working day, in person, with the team

Compensation

• Top-of-market base, calibrated to live demonstration of capability
• Meaningful equity reflecting the seniority and scope of the role
• Direct, no-gatekeeper access to the founders
• Full ownership of architecture choices within your scope from day one

To Apply

Resumes will only be reviewed if the application form is completed.

> Application form: https://forms.gle/ciFTj8iS6j3TLVNt7

The form takes 5 minutes. Submitting a CV outside the form will not be reviewed. Shortlisted candidates will be contacted within 48 hours of submission.

Steps AI · stepsai.co · Hyderabad, India