New opportunity

Machine Learning Engineer

HexanixOne · Ahmedabad, India

About this role

HexanixOne seeks a Machine Learning Engineer to build and deploy production ML systems for enterprise clients. Projects range from LLM fine-tuning to computer vision applications, with cross-functional collaboration to deliver AI solutions at scale. The posting lists the location as Ahmedabad / Remote and offers a hybrid work model with flexible hours.

Responsibilities

Design and implement ML models for different use cases; fine-tune and deploy LLMs for enterprise applications.
Build MLOps pipelines for training and deployment, optimize model performance and inference latency, and monitor production models to guide retraining.
Translate product requirements into ML solutions, evaluate new techniques and frameworks, and contribute to ML infrastructure and tooling.

Required qualifications

At least 4 years of ML engineering experience; strong Python skills and experience with TensorFlow and PyTorch.
Experience with LLMs such as GPT-4, Claude and Llama, including fine-tuning; knowledge of MLOps tools and practices.
Understanding of statistics and ML algorithms, experience deploying models to production at scale, and strong software engineering fundamentals.
An MS or BS in Computer Science, ML or a related field.

Preferred qualifications include a related PhD; AWS SageMaker or Azure ML experience; published research or contributions to ML frameworks; computer vision or NLP experience; knowledge of distributed training and optimization; and a background in deep learning and neural architecture search.

The posting mentions competitive compensation without a specific salary, professional development, fully sponsored Microsoft certifications, team-building activities, and a generous PTO and holiday schedule.

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

Skills for this role

PythonTensorFlowPyTorchMachine LearningLarge Language ModelsLLM fine-tuningGPT-4ClaudeLlamaMLOpsStatisticsML algorithmsModel deploymentModel monitoringInference optimizationSoftware engineeringAWS SageMakerAzure MLComputer VisionNatural Language ProcessingDistributed trainingDeep LearningNeural architecture search

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

Production ML delivery
The role calls for designing models for varied use cases and deploying them at scale. Prepare to explain an end-to-end production system, including model choices and software engineering decisions.
Enterprise LLM applications
Fine-tuning and deploying LLMs is an explicit responsibility. Prepare to discuss how you would evaluate a fine-tuned model against an enterprise use case before deployment.
MLOps and model lifecycle
The posting calls for training and deployment pipelines, production monitoring, and retraining strategies. Prepare a concrete account of how those stages connect and how monitoring would inform retraining.
Product alignment and technical evaluation
The engineer works with product teams to translate requirements into ML solutions, evaluates new techniques, and contributes to infrastructure and tooling. Prepare to explain how requirements shape a technical choice and how you would test an alternative.

Likely challenges to think through

Model quality versus inference latency
A likely trade-off is improving model performance while meeting the stated goal of reducing inference latency. Prepare to describe the metrics, comparisons, and deployment constraints you would use to make that choice; the posting does not specify latency targets.
Varied use cases without one default approach
Projects range from LLM fine-tuning to computer vision. A likely difficulty is deciding when an existing technique fits and when a different approach is warranted. Prepare a decision framework based on the use case, data, and evaluation results rather than assuming one model family applies throughout.
Retraining without disrupting production
Because the role includes monitoring and retraining, a likely challenge is distinguishing meaningful performance decline from routine variation while keeping deployments reliable. Prepare to explain what evidence would trigger retraining and how you would validate a replacement model before release.

Practice your answers

Technical practice

How would you turn a product requirement into an ML approach?

What to coverThe role involves translating product requirements into ML solutions. Suggested preparation: explain how you would define the target, choose a baseline and evaluation metric, and use statistical evidence to decide whether the model meets the requirement.

When would you fine-tune an LLM for an enterprise application?

What to coverFine-tuning and deploying LLMs are stated responsibilities. Suggested preparation: compare fine-tuning with simpler approaches, and discuss data quality, evaluation, cost and deployment constraints without assuming every named model supports the same options.

How would you design and evaluate a computer vision model?

What to coverThe page mentions computer vision projects and lists CV experience as nice to have. Suggested preparation: describe a task-appropriate baseline, data and label checks, error analysis, and metrics tied to the intended use.

What would an MLOps pipeline cover from training through deployment?

What to coverThe role calls for training and deployment pipelines. Suggested preparation: cover reproducible data and model versions, validation before release, and a way to trace a deployed model back to its training inputs.

How would you deploy an ML model reliably at scale?

What to coverProduction deployment at scale and software engineering fundamentals are required. Suggested preparation: discuss serving interfaces, testing, capacity, failure handling and rollout trade-offs for a model-backed service.

How would you reduce inference latency without losing necessary model quality?

What to coverOptimizing performance and reducing inference latency are stated duties. Suggested preparation: explain how you would profile the bottleneck, compare optimization options, and measure quality, throughput and latency together.

How would you decide when a production model needs retraining?

What to coverThe role includes monitoring model performance and implementing retraining strategies. Suggested preparation: distinguish service health from model quality, explain what signals you would track, and describe how you would validate a replacement before release.

How would you structure and test Python model code across ML frameworks?

What to coverThe posting requires strong Python skills and experience with TensorFlow and PyTorch. Suggested preparation: discuss maintainable training and inference code, tests for data and model behavior, and how framework choice affects implementation.

How would you evaluate a new ML technique or framework?

What to coverResearching and evaluating techniques and frameworks is a stated responsibility. Suggested preparation: propose a comparison against a relevant baseline using consistent data, metrics and resource costs, then explain what evidence would justify adoption.

What shared ML tooling would you build first, and why?

What to coverThe role includes contributing to ML infrastructure and tooling. Suggested preparation: identify a recurring training or deployment problem, describe a small reusable solution, and explain how you would assess whether it improves reliability or developer effort.

Behavioural practice

Tell me about a time you clarified an ambiguous product request.

What to coverThe role involves working with product teams to translate requirements into ML solutions. Suggested preparation: use a real situation/action/result example showing the questions you asked, the scope you agreed on and the outcome.

Describe a cross-functional disagreement about an ML trade-off.

What to coverThe page describes collaboration across teams on enterprise AI projects. Suggested preparation: give a real situation/action/result example that shows how you explained the trade-off, incorporated other perspectives and reached a decision.

Tell me about a production model problem you helped resolve.

What to coverDeployment and production monitoring are central to the role. Suggested preparation: use a real situation/action/result example covering how the issue was detected, what you did and how you checked the outcome.

Describe a time you assessed a promising new ML approach.

What to coverThe role includes evaluating new techniques and frameworks. Suggested preparation: give a real situation/action/result example showing your comparison criteria, evidence and decision, including why an approach was or was not adopted.

Tell me about a reusable improvement you made to an ML workflow.

What to coverThe posting calls for contributions to ML infrastructure and tooling. Suggested preparation: use a real situation/action/result example that identifies the recurring problem, your contribution and its observable effect on the workflow.

Sharpen your knowledge

Topics to prepare

Statistics and ML algorithms
Required: a solid understanding of both. As a revision exercise, compare two algorithms for a sample prediction problem, explaining assumptions, evaluation metrics, and how you would detect overfitting.
LLM fine-tuning decisions
The role calls for LLM experience and fine-tuning; GPT-4, Claude, and Llama are named. As a revision exercise, outline when you would fine-tune rather than use an existing model as-is, and how you would evaluate the result. Do not assume every named model is fine-tunable.
MLOps and retraining
The posting requires MLOps knowledge and calls for training, deployment, monitoring, and retraining. As a revision exercise, sketch that lifecycle and identify what evidence would prompt a retraining decision.
Production performance trade-offs
The role involves deployment at scale, model performance, and inference latency. As a revision exercise, explain how you would compare model quality, latency, and operational cost before choosing a deployment approach.
Turning product needs into ML work
The posting explicitly calls for collaboration with product teams. As a revision exercise, take a hypothetical enterprise request and define the ML objective, success measures, and questions you would clarify before modeling.

Skills to revise

Implement models in Python frameworks
Strong Python skills and experience with TensorFlow and PyTorch are required. As a practice exercise, implement and evaluate the same small model in each framework, then explain the code and results.
Practice an LLM fine-tuning workflow
Fine-tuning and deploying LLMs are stated responsibilities. As a suggested exercise, prepare a small task dataset, fine-tune a Llama model, compare it with a baseline, and describe how you would deploy it.
Build a training-to-deployment workflow
The role calls for MLOps pipelines and production deployment. As a practice exercise, connect data preparation, training, evaluation, and deployment steps, with checks that prevent an underperforming model from being released.
Measure and reduce inference latency
Reducing inference latency is an explicit responsibility. As a practice exercise, measure latency for a sample model, change one implementation choice, and report both the latency change and any effect on output quality.
Monitor a deployed model
The posting calls for monitoring performance and implementing retraining strategies. As a practice exercise, define a small set of production measurements, specify a trigger for investigation, and describe how you would validate a retrained model.
Explain an ML solution to a product partner
Cross-functional collaboration is part of the role. As a suggested exercise, present a short solution plan for a hypothetical product request, covering the proposed model, evaluation criteria, deployment considerations, and unresolved questions.

Your preparation checklist

Map the required qualifications
Prepare a factual one-page summary of your ML engineering experience, degree, Python work, and use of TensorFlow or PyTorch. The posting specifies 4+ years of experience and a related BS or MS; leave any unsupported item unclaimed.
Rehearse a production ML walkthrough
Choose a real project, if available, and explain the problem, data, model choice, evaluation, deployment, scale, and your contribution. The role centers on production ML systems for enterprise clients.
Prepare an LLM application example
Be ready to discuss an actual LLM project or a clearly labeled proposed approach to fine-tuning and deploying an enterprise application. Explain evaluation and trade-offs; the posting names GPT-4, Claude, and Llama, but does not require the same fine-tuning method for each.
Sketch a training-to-production pipeline
Draw a design covering training, validation, deployment, monitoring, and retraining triggers. Mark which parts reflect your experience and which are proposed; the role explicitly includes MLOps pipelines and production monitoring.
Quantify performance and latency trade-offs
Prepare any available measurements showing model quality, inference latency, and the effect of an optimization. If you lack a measured example, outline how you would benchmark these for a proposed solution.
Practice translating a product request
Work through a hypothetical enterprise use case: clarify the objective with product partners, select success metrics, compare ML approaches, and identify deployment constraints. This exercises the cross-functional responsibility stated in the posting.
Write focused questions for the employer
Ask how teams choose among LLM and computer vision projects, what production monitoring and retraining look like, and where this role would contribute to ML infrastructure. These are questions, not assumptions about HexanixOne’s processes.

Availability last checked: 5 Nov 2026. 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…