New opportunity

Python AI / Machine Learning Developer

ExcelPTP Solutions · Ahmedabad, India

About this role

ExcelPTP Solutions seeks a Python AI / Machine Learning Developer in Navrangpura, Ahmedabad, to build intelligent applications for live international projects. The full-time role has two openings and working hours of Monday to Friday, 9:30 AM to 7:00 PM.

Responsibilities

Develop AI and machine learning applications in Python; build, train, evaluate and optimize models.
Work with LLMs, generative AI, prompt engineering and AI APIs. Preprocess data, engineer features and evaluate models.
Develop REST APIs using FastAPI, Flask or Django, and integrate AI models into web applications and enterprise systems.
Write clean, scalable, documented code; collaborate with developers, understand business requirements and keep current with AI, ML and Python technologies.

Requirements

The posting lists 1+ years of total experience, 1–4 years of relevant experience and, in the detailed description, 12 months–4 years of experience. It separately specifies 6 months–2 years of Python, AI or machine learning development experience.
Strong Python, object-oriented programming, data structures and algorithms; experience with NumPy, Pandas, Scikit-learn, TensorFlow or PyTorch; knowledge of generative AI, LLMs, prompt engineering, LangChain or OpenAI APIs.
Experience with FastAPI, Flask or Django; knowledge of SQL, REST APIs, JSON and Git/GitHub. Effective use of AI development tools such as ChatGPT, GitHub Copilot, Cursor AI, Claude or Gemini. Analytical thinking, problem-solving, communication, teamwork and deadline management are required.
The listing says undergraduates or any graduates may apply.

Preferred

B.E., B.Tech, MCA, M.Sc. in IT, CS, AI or Data Science, BCA or an equivalent qualification. Relevant certifications are an advantage. Docker, AWS, Azure AI, Google Cloud AI or MLOps experience is preferred; knowledge of vector databases, RAG, AI agents and AI automation is a plus. The posting mentions mentorship, certification support and performance-based career growth.

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

Skills for this role

PythonMachine LearningGenerative AILarge Language ModelsPrompt EngineeringAI APIsModel EvaluationData PreprocessingFeature EngineeringREST APIsFastAPIFlaskDjangoObject-Oriented ProgrammingData StructuresAlgorithmsNumPyPandasScikit-learnTensorFlowPyTorchLangChainOpenAI APIsSQLJSONGitGitHubChatGPTGitHub CopilotCursor AI tools not included under standardized name? Correct: Cursor AI remains

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

Build and evaluate ML models
The role calls for preprocessing data, engineering features, and building, training, evaluating, and optimizing models. Prepare to explain an end-to-end Python workflow and how evaluation results guide changes.
Develop LLM-powered applications
The posting explicitly includes generative AI, LLMs, prompt engineering, and AI APIs. Prepare to discuss how you would design prompts, use an API, and assess whether outputs meet an application’s needs.
Deliver AI through web APIs
The developer will build REST APIs with FastAPI, Flask, or Django and integrate models into web applications and enterprise systems. Prepare to explain the path from model input and output to a documented API endpoint.
Turn requirements into maintainable code
The posting emphasizes understanding business requirements, collaborating with developers, and writing clean, scalable, documented code while meeting deadlines. Prepare to describe how you would clarify a requirement, structure the Python code, and communicate progress.

Likely challenges to think through

Balancing model quality with application constraints
A likely challenge, not a stated employer problem, is that a model that evaluates well may still be difficult to integrate or serve through a REST API. Prepare to discuss trade-offs among output quality, response time, and maintainable integration.
Judging variable LLM outputs
Because the role includes prompts, LLMs, and AI APIs, consistent output quality may be harder to establish than for a conventional model. Prepare a way to compare prompt or API changes against representative examples and explain their limitations.
Prioritizing across AI and software tasks
The posting combines model work, API development, collaboration, and deadlines. A plausible difficulty is deciding what to deliver first when requirements evolve; prepare to explain how you would clarify the goal, identify dependencies, and communicate trade-offs.

Practice your answers

Technical practice

How would you structure a Python application that trains a model and serves predictions?

What to coverSuggested answer focus: explain how you would separate data preparation, training, inference, and API code using clear interfaces and object-oriented design where useful. Discuss how you would keep the code documented, testable, and maintainable.

How would you clean and inspect a dataset before training?

What to coverSuggested answer focus: describe checks for missing values, duplicates, inconsistent types, and unexpected distributions using tools such as Pandas or NumPy. Explain how the findings would guide preprocessing rather than assuming one cleaning rule fits every dataset.

How would you engineer features without leaking information into evaluation?

What to coverSuggested answer focus: explain how you would choose features relevant to the business goal and fit preprocessing steps on training data only. Discuss how leakage can make evaluation results misleading.

How would you train and evaluate a machine-learning model?

What to coverSuggested answer focus: describe a baseline, an appropriate data split, and metrics chosen for the task and business objective. Explain how you would interpret errors, not just report a single score.

What would you try if a model performed well on training data but poorly on unseen data?

What to coverSuggested answer focus: discuss checks for overfitting, data quality, and differences between training and evaluation data. Compare possible changes such as simpler models, regularization, or better features, and say how you would test whether they help.

How would you design and evaluate a prompt for an LLM-powered feature?

What to coverSuggested answer focus: explain how you would specify the task, context, and expected output, then test the prompt against representative and difficult inputs. Discuss trade-offs between output quality, consistency, latency, and API cost.

How would you integrate an AI API into a Python application?

What to coverSuggested answer focus: describe request and response handling, configuration, error handling, and validation of outputs before the application uses them. Explain what you would do when the external API is slow, unavailable, or returns an unsuitable response.

How would you expose a model through a REST API for a web application?

What to coverSuggested answer focus: outline a prediction endpoint using FastAPI, Flask, or Django, including JSON input validation and a clear response format. Discuss how you would handle model loading, errors, and changes to the API contract.

How would you trace a faulty prediction from a web request back to its source?

What to coverSuggested answer focus: walk through checking the JSON payload, application logic, preprocessing, model version, and underlying SQL data where relevant. Explain how tests, readable code, and Git history could help isolate a change.

When might retrieval-augmented generation be useful alongside an LLM?

What to coverSuggested preparation for a topic listed as a plus, not a core requirement: explain how retrieval from a vector database could supply relevant context before generation. Compare that approach with prompting an LLM alone, including the need to evaluate retrieval quality and generated answers.

Behavioural practice

Tell me about a time you clarified an unclear business requirement before building a solution.

What to coverSuggested answer focus: use a real situation, the questions you asked, the action you took, and the result. Explain how the clarified goal affected your technical choices.

Describe a time you collaborated with developers to integrate or improve a feature.

What to coverSuggested answer focus: give a real example covering your role, how you communicated interfaces or changes, and the result. Include how you handled a differing technical view if one arose.

Tell me about a time you diagnosed a difficult data, model, or application problem.

What to coverSuggested answer focus: describe the symptoms, your investigation, the evidence behind your fix, and the result. Distinguish what you verified from what you initially suspected.

Describe a time you had to prioritize competing tasks against a deadline.

What to coverSuggested answer focus: use a real situation to explain how you assessed urgency and impact, communicated trade-offs, and acted. State the outcome and anything you would change next time.

Tell me about a time you learned or used a new AI or Python tool for development.

What to coverSuggested answer focus: describe why you chose the tool, how you checked its output, and what resulted. If an AI coding assistant was involved, explain how you kept responsibility for code quality and correctness.

Sharpen your knowledge

Topics to prepare

Python design and algorithmic reasoning
Review OOP, data structures, and time complexity by explaining two ways to organize a model-serving component and comparing their trade-offs. The posting explicitly requires strong Python, OOP, data structures, and algorithms.
Machine learning lifecycle
Trace a sample problem from preprocessing and feature engineering through training, evaluation, and optimization; explain how you would spot data leakage. These stages are explicit responsibilities.
Generative AI and LLM behavior
Compare prompts for one task, define what a useful response would look like, and explain how you would evaluate failures. The role explicitly involves LLMs, generative AI, prompt engineering, and AI APIs.
Model integration through REST
Sketch a JSON request and response for an AI-powered web feature, then explain validation and error handling. The posting calls for REST APIs and integration into web applications and enterprise systems.

Skills to revise

Prepare data and evaluate a model
Using Pandas and Scikit-learn, clean a small dataset, create features, train a baseline, and record evaluation results. Both libraries are listed among alternatives, and the workflow matches stated responsibilities.
Build a Python model endpoint
Using one listed framework—FastAPI, Flask, or Django—implement a REST endpoint that accepts JSON, validates input, and returns a model result. This practices the advertised API-development and integration work.
Work with an LLM API
Try a small prompt-driven interaction using a listed option such as an OpenAI API; compare outputs for two prompt versions and document a failure case. This practices the posting’s AI API and prompt-engineering work.
Use SQL, JSON, and Git together
Query sample records with SQL, shape the result as JSON, and commit the changes with a descriptive Git message. The posting explicitly names SQL, JSON, and Git/GitHub.
Review AI-assisted code critically
Use one listed AI development tool to suggest a Python change, then test, correct, and explain the final code yourself. Effective use of such tools and clean, documented code are stated requirements.

Your preparation checklist

Prepare a machine-learning project walkthrough
The role involves preprocessing data, engineering features, and building, evaluating, and optimizing models. Prepare a walkthrough of work you can substantiate, including the problem, data choices, evaluation metric, results, and one improvement; if you lack a completed project, use a clearly labeled practice example.
Practice an LLM integration example
The posting names LLMs, generative AI, prompt engineering, and AI APIs. Prepare a small example or design sketch showing a prompt, API interaction, how you would assess output quality, and a limitation. Label any proposed approach as practice rather than job experience.
Sketch a model-backed REST API
As a preparation exercise, outline a Python endpoint using FastAPI, Flask, or Django that accepts JSON and returns a model result. Bring a request-and-response example and explain validation, error handling, and how the model would connect to a web application.
Collect evidence of Python development
Select code you are permitted to share that demonstrates Python, object-oriented design, data structures, SQL or JSON, and Git usage where applicable. Be ready to explain your own decisions, tests, documentation, and any AI development tools you used.
Rehearse a business-requirements discussion
The role calls for translating business requirements into AI solutions and collaborating with developers. Practice clarifying a sample request, choosing an evaluation measure, and explaining trade-offs, priorities, and a delivery plan without assuming the company's process.
Draft questions about the work
Ask which types of live AI projects the opening would support, how models are evaluated and integrated into applications, and which listed technologies are used day to day. Ask whether preferred areas such as RAG, cloud AI, or MLOps matter for this opening.
Prepare the requested application details
The posting asks for a current CV plus total and relevant experience, reason for a job change, current and expected salary, and availability to join. Have accurate answers ready, and note any questions you have about the stated Navrangpura location and Monday–Friday hours.

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…