Role: Applied AI Engineer
Job type: FT with Highspring
Shift: 2 to 11 pm IST
Location: 100% remote
Applied AI Engineer
Role Overview
We are looking for an Applied AI Engineer to design production-grade autonomous agentic systems while maintaining and optimizing our existing traditional machine learning pipelines. In this role, you will balance cutting-edge LLM orchestration with the upkeep of algorithm-heavy legacy workflows, including OCR systems.
Key Responsibilities
- Develop Agentic Workflows: Build autonomous multi-agent systems, cognitive loops, and reasoning workflows utilizing tool execution and self-correction.
- Maintain Traditional ML & OCR: Support, optimize, and scale existing algorithmic pipelines, with a specific focus on Optical Character Recognition (OCR) and heavy backend processing algorithms.
- Production Deployment: Ship both generative and traditional AI prototypes into scalable, low-latency production environments.
- State & Memory Management: Design robust state machines to handle agent memory, context, and multi-step persistence.
- Optimize & Evaluate: Implement frameworks to monitor model behavior, mitigate hallucinations, and optimize compute/token spend.
Technical Qualifications
Core Fundamentals
- Software Engineering: Strong command of Object-Oriented Programming (OOP), data structures, and clean code architecture.
- Machine Learning: Solid grounding in core ML fundamentals, including supervised/unsupervised learning algorithms, and traditional computer vision/OCR techniques.
Required AI & Python Expertise
- Python: 4+ years of professional experience (strong focus on asynchronous programming, Pydantic, and FastAPI).
- Agentic AI Frameworks: Hands-on experience with LangGraph, Strands SDK, CrewAI, or building custom agent loops.
- LLM Orchestration: Mastery of advanced prompt engineering, function calling, and complex Retrieval-Augmented Generation (RAG) patterns.
- MCP server and client implementations
Preferred
- Implementing cloud-native solutions on AWS and hands-on exposure of building production workloads with: AWS Bedrock: Provisioning, fine-tuning, and orchestrating foundational models. AWS Lambda: Building event-driven architecture and executing isolated agent tools. AWS DynamoDB: Managing session persistence, agent memory states, and fast transactional storage.
- Vector Architecture: Familiarity with vector databases (e.g., OpenSearch, FAISS) for knowledge retrieval.
- Experience with API and Micro-services implementations
Notes from the hiring manager
He is looking for a proven engineer whose deliverables we can vouch for. Experience with modern AWS stack and traditional ML is a must. Hiring an unknown entity with unverified credentials may cause a lot of delay in ramp up. He mentioned it can take up to 2 years before they start delivering something meaningful. Projects include enhancing existing product line and building new ones. He is not looking for folks who have done only Agentic AI work. He is looking for hands on experience with traditional Machine Language and Optical Character Recognition (OCR) experience.