New opportunity

Sr ML & AI Engineer

HERE Technologies · Mumbai, India

About this role

As a Senior ML & AI Engineer at HERE Technologies, you will design, develop, and deploy intelligent systems that combine Generative AI, Agentic AI, machine learning, and advanced retrieval techniques to solve complex business and spatial challenges. You will architect AI-first systems, build representation and retrieval layers, and lead the development of agentic solutions for task orchestration and human-agent collaboration. Responsibilities include delivering end-to-end Generative AI applications, developing multimodal vision-language systems, and applying deep learning techniques to real-world tasks. You will collaborate across teams to translate domain problems into scalable, production-ready AI solutions while adhering to responsible AI principles. Qualifications include 6+ years of experience in ML/AI engineering with a track record of shipping production systems, a strong foundation in representation learning, vector similarity search, and retrieval-augmented approaches, and proficiency in Python and modern ML frameworks like PyTorch or TensorFlow. Preferred qualifications include experience with agentic systems (LangGraph, CrewAI, AutoGen), multimodal models, and geospatial data processing. Candidates with a Master's degree should have 2–7 years of experience, while Ph.D. holders should have 2–5 years of relevant experience. The role involves mentoring team members, contributing to innovation through patents or publications, and staying at the forefront of AI research.

For full details about this role, click “Apply on company site” to view the employer’s original posting.

Skills for this role

PythonPyTorchTensorFlowLarge Language Models (LLMs)Generative AIAgentic AILangGraphCrewAIAutoGenDeep LearningComputer VisionNatural Language Processing (NLP)Vector DatabasesSparkGeospatial DataMLOpsRetrieval-Augmented Generation (RAG)Representation LearningTransformer ArchitecturesC++Java

Checking your profile…

Preparation suggestions, not the employer’s actual interview questions. Always check the original posting for current requirements.

Understand the role

What matters most

Architecting Scalable AI-First Systems
The role requires designing and deploying end-to-end intelligent systems that integrate Generative AI, Agentic AI, and classical machine learning. You must demonstrate the ability to own ambiguous, end-to-end problems and make independent architectural decisions that translate domain-specific business challenges into production-ready solutions.
Developing Advanced Retrieval and Representation Layers
A core priority is building the infrastructure that powers intelligent systems, specifically focusing on embedding models, unified latent-space representations, and vector similarity search. You will be expected to implement retrieval-augmented generation (RAG) at scale to support complex spatial and business reasoning.
Orchestrating Agentic AI and Multimodal Workflows
You will lead the development of agentic solutions for task orchestration, memory management, and tool invocation using frameworks like LangGraph or AutoGen. Additionally, you must be capable of building multimodal systems that jointly encode imagery with text and structured attributes for semantic reasoning.
Applying Geospatial Reasoning at Scale
Given the company's focus on location data, you must prioritize the application of ML to spatial data structures, indexing, and entity resolution. You should be prepared to demonstrate how your AI models handle spatial reasoning and large-scale geospatial datasets.

Likely challenges to think through

Balancing Innovation with Production Stability
The role demands staying at the forefront of AI research while ensuring systems are production-ready and explainable. Prepare by discussing how you manage the trade-off between implementing cutting-edge agentic frameworks and maintaining the reliability, confidence-gating, and MLOps standards required for enterprise-grade applications.
Integrating Heterogeneous Data Signals
You will be tasked with creating unified representations across diverse signals, including imagery, text, and structured spatial data. Prepare to explain your approach to feature engineering and model architecture when dealing with the high dimensionality and noise inherent in multimodal geospatial datasets.

Practice your answers

Technical practice

Architecting Agentic AI Workflows

What to coverExplain how you would design a multi-agent system using frameworks like LangGraph or CrewAI for a complex task. Focus on how you manage state, handle tool invocation, and ensure reliable communication between agents.

Retrieval-Augmented Generation (RAG) at Scale

What to coverDescribe your approach to building a robust retrieval layer for LLMs. Discuss the trade-offs between different embedding models, vector database indexing strategies, and how you handle retrieval quality in large-scale production environments.

Multimodal Vision-Language Integration

What to coverHow do you approach jointly encoding imagery and structured text data? Discuss the challenges of aligning latent spaces and how you would evaluate the performance of a vision-language model in a real-world spatial reasoning task.

Geospatial Data Processing and Indexing

What to coverGiven the focus on spatial reasoning, explain how you handle large-scale geospatial datasets. Discuss your experience with spatial indexing techniques or entity resolution/conflation over spatial records.

Confidence Calibration in Autonomous Systems

What to coverHow do you design confidence-gating and escalation logic for AI systems? Explain how you incorporate human-in-the-loop or active-learning workflows to improve system reliability and safety.

Productionizing Deep Learning Models

What to coverDescribe your MLOps pipeline for deploying deep learning models. Focus on how you manage data versioning, model evaluation, and monitoring for drift in a production environment.

Optimization of LLM Inference

What to coverWhat techniques do you use to optimize LLM performance for production applications? Address strategies such as prompt engineering, fine-tuning, or quantization to balance latency and accuracy.

Representation Learning Fundamentals

What to coverExplain the importance of representation learning in modern AI systems. Discuss how you design embeddings that capture meaningful relationships across heterogeneous data signals.

Behavioural practice

Leading Through Ambiguity

What to coverDescribe a time you were tasked with solving an ambiguous, end-to-end AI problem. Explain how you defined the scope, made architectural decisions, and ensured the final solution met business needs.

Mentoring and Technical Leadership

What to coverProvide an example of how you have mentored junior engineers or data scientists. Focus on how you guided them through research translation, evaluation frameworks, or best practices for AI development.

Cross-Functional Collaboration

What to coverTell us about a time you had to translate a complex domain problem into a scalable AI solution while working with non-technical stakeholders. How did you ensure the solution was explainable and aligned with responsible AI principles?

Driving Innovation and Impact

What to coverDescribe a situation where you identified an opportunity for innovation, such as a patent or a new research implementation. How did you balance this experimental work with the need to deliver production-ready systems?

Sharpen your knowledge

Topics to prepare

Agentic AI Architectures
Review the design patterns for multi-agent systems, specifically focusing on task orchestration, memory management, and human-in-the-loop workflows. This is highly relevant as the role requires leading the development of agentic solutions using frameworks like LangGraph or AutoGen to solve complex spatial challenges.
Retrieval-Augmented Generation (RAG) and Vector Search
Study the mechanics of building representation layers, including embedding models and vector similarity search at scale. This is a core requirement for the position, as you will be architecting systems that combine advanced retrieval techniques with LLMs to process heterogeneous signals.
Multimodal Vision-Language Systems
Explore techniques for jointly encoding imagery with text and structured attributes. Given HERE Technologies' focus on location data, understanding how to integrate visual data with semantic reasoning is essential for the role's emphasis on spatial reasoning.

Skills to revise

Agentic Framework Implementation
Practice building a multi-agent workflow using LangGraph or CrewAI to automate a multi-step task. This directly addresses the preferred qualification for experience with agentic systems and task orchestration.
Geospatial Data Processing
Familiarize yourself with spatial indexing and entity resolution techniques. Since the role centers on spatial reasoning at scale, demonstrating an ability to handle geospatial data structures will be a significant advantage.
Production-Grade Deep Learning with PyTorch/TensorFlow
Develop a pipeline that includes model training, evaluation, and deployment using PyTorch or TensorFlow. The role requires a track record of shipping production systems, not just research prototypes, so focus on MLOps practices and scalable inference.

Your preparation checklist

Prepare a production-scale AI project walkthrough
Select a past project where you deployed a deep learning or LLM-based system. Prepare to discuss the end-to-end architecture, specifically focusing on how you handled data pipelines, MLOps practices, and the transition from prototype to production-ready code.
Demonstrate expertise in agentic and retrieval systems
Be ready to explain your experience with frameworks like LangGraph, CrewAI, or AutoGen. Prepare a brief technical summary of how you have implemented RAG or vector similarity search to solve complex retrieval challenges at scale.
Connect geospatial data to AI solutions
Since HERE Technologies focuses on spatial reasoning, prepare examples of how you have worked with spatial indexing, entity resolution, or geospatial data structures. If you lack direct experience, research how multimodal models can be applied to spatial data.
Formulate technical questions for the team
Prepare questions regarding their current MLOps stack, how they manage human-in-the-loop confidence calibration for autonomous systems, and how they balance research-driven innovation with production stability.

Availability has not yet been verified. Status can change after a check; confirm on the employer’s website. Report an issue · Our checking policy

KEEP EXPLORING

Similar AI jobs

Related skills and specializations in India. Matched to this role, not your profile.

Explore more jobs

Finding similar opportunities…