Sr. AI Application Dev

About ETS Labs

ETS Labs, an Etech Global Services Company, is a technology-

driven organization building enterprise-grade AI applications,

analytics platforms, and cloud solutions for global clients across

contact center, healthcare, and financial services domains. We are

at the forefront of AI-powered product engineering — delivering

conversational AI, agentic automation, RAG-based knowledge

platforms, and real-time analytics at scale.

Role Overview

We are looking for a Senior AI Application Developer & Architect

who specializes in building production-grade AI-powered software

applications. The primary focus of this role is end-to-end

development of GenAI applications — including conversational AI

systems, agentic workflows, RAG pipelines, LLM-integrated APIs,

and real-time AI services. This is an application development role

where AI is the core product layer, not a research or model-training

position. The ideal candidate thinks like a software engineer first

and uses LLMs, agentic frameworks, and cloud AI services as the

primary building blocks.

Key Responsibilities

AI Application Development (Primary Focus)

· Design and build end-to-end AI-powered applications —

conversational chatbots, agentic assistants, document intelligence

systems, and real-time AI analytics platforms.

· Develop LangGraph-based agentic workflows with multi-step

reasoning, tool orchestration, HITL approval gates, and crash-

recovery state persistence.

· Build and integrate LLM APIs (AWS Bedrock, OpenAI, Groq,

Gemini, Ollama) into production application backends with

pluggable provider abstraction.

· Develop real-time AI features using FastAPI and WebSocket

streaming, delivering token-by-token LLM responses to end users.

· Build Text-to-SQL engines, automated data visualization

pipelines, and AI-driven analytics features within application layers.

· Integrate multimodal AI capabilities — OCR, document

parsing, image understanding — into application workflows where

required.

RAG & Knowledge Retrieval Systems

· Build production-grade RAG pipelines integrating vector

databases (Weaviate, Pinecone, OpenSearch) with hybrid dense +

BM25 retrieval.

· Implement query transformation, reranking (Cohere,

CrossEncoder), and LLM-based citation validation within application

flows.

· Design multi-tenant document ingestion pipelines with per-

user isolation, lifecycle tracking, and scheduled processing.

· Develop knowledge base chatbot applications with multi-turn

conversational memory and sliding window context compaction.

Team Leadership & Solution Architecture

· Lead a small team of 2–3 AI developers, conducting code

reviews, architecture walkthroughs, and delivery planning.

· Create solution architecture diagrams covering application,

integration, data flow, cloud, and security layers.

· Act as the technical owner for AI application delivery — from

requirements to production deployment.

· Translate business requirements and client use cases into AI

application designs and implementation roadmaps.

Backend Engineering & Cloud Deployment

· Build scalable backend services in Python (FastAPI) with

async concurrency, task queuing (Celery + Redis), and scheduled

processing (APScheduler).

· Implement RBAC systems, multi-tenant data isolation, and API

security patterns within AI application architectures.

· Deploy AI applications on AWS (ECS Fargate, Lambda) via

CI/CD pipelines with container orchestration and secrets

management.

· Apply performance engineering practices: async circuit

breakers, retry logic, connection pooling, and memory optimization

for production AI workloads.

Responsible AI & Quality

· Implement AI guardrails, prompt injection prevention, output

validation, and content moderation within application pipelines.

· Build PII detection, audit logging, traceability, and

explainability features for compliance-sensitive AI applications.

· Write unit and integration tests for AI application components,

ensuring reliability of LLM-integrated workflows.

Required Skills & Expertise

AI Application Development (Core — Must Have)

· LangGraph, LangChain — agentic workflow design, state

machines, conditional routing, tool nodes

· LLM API integration — AWS Bedrock (Claude, Llama, Mistral,

Titan), OpenAI GPT-4o, Gemini, Groq, Ollama

· Prompt Engineering — structured prompts, output formatting,

few-shot design, chain-of-thought reasoning

· RAG pipeline development — document ingestion, chunking,

embedding, hybrid retrieval, reranking, generation

· Conversational AI — multi-turn memory, session

management, context compaction, streaming responses

· HITL workflow design — approval gates, escalation flows,

human override mechanisms

Backend & API Development

· Python — FastAPI, asyncio, REST API design, WebSocket

streaming

· Task queuing — Celery, Redis; Scheduling — APScheduler

· Databases — MongoDB, PostgreSQL, SQLite for application

data and lifecycle tracking

· Authentication & Authorization — JWT, RBAC, OAuth, multi-

tenant patterns

Vector Databases & Search

· Weaviate (multi-tenant), Pinecone (namespace isolation),

OpenSearch — production deployment experience

· Hybrid retrieval: BM25 + dense vector, reranking with

CrossEncoder or Cohere

· Embedding models: Amazon Titan Embed v2, OpenAI Ada,

local sentence transformers

Cloud & DevOps

· AWS — Bedrock, ECS (Fargate), Lambda, S3, Secrets

Manager, CloudWatch

· Docker, Kubernetes basics, GitLab / GitHub CI/CD pipelines

· Infrastructure as Code awareness (Terraform /

CloudFormation) is a plus

ML & NLP Awareness (Good to Have — Not Primary)

· Basic understanding of NLP concepts: tokenization,

embeddings, text classification — sufficient to work with pre-trained

models via APIs.

· Familiarity with Hugging Face model hub for accessing pre-

trained models (BERT, sentence transformers) when needed in

application pipelines.

· Understanding of when to use fine-tuned models vs. prompt

engineering vs. RAG — to make the right architectural choice.

· Experience with traditional ML frameworks (TensorFlow, scikit-

learn) is a plus but not required for this role.

Responsible AI

· AI guardrails, output validation, prompt injection prevention,

and content moderation

· PII detection, compliance monitoring, audit trails, and

traceability in AI application outputs

· Multi-tenant data isolation and security-aware AI application

design

Preferred Qualifications

· B.Tech / M.Tech / BCA / MCA in Computer Science, Software

Engineering, AI/ML, or equivalent.

· 5+ years of software development experience with at least 3

years in AI application development.

· Proven track record of delivering 3+ production AI applications

(chatbots, agentic systems, RAG platforms).

· Demonstrated experience with LangGraph or similar agentic

orchestration frameworks.

· Exposure to contact center AI, document intelligence, or

enterprise analytics AI platforms.

· Strong API design, code quality, and software engineering

fundamentals.

· AWS certifications (Developer, Solutions Architect) are a plus;

ML Specialty not required.

Nice to Have

· Experience with MCP (Model Context Protocol) server

development and tool orchestration.

· Voice bot or speech-to-text/speech translation pipeline

development.

· Exposure to multimodal AI: OCR, document parsing, image

understanding within application workflows.

· Knowledge of FedRAMP, HIPAA, SOC 2, or regulated-industry

AI compliance requirements.

· Experience with streaming front-end integration — React or Angular consuming WebSocket AI responses.