Machine Learning – Internship (WFH)
DSights Consultancy Services Private Limited is hiring a Machine Learning Engineer Intern for a paid, work-from-home opportunity focused on structured and tabular data. The four-month internship advertises ₹20,000–30,000 per month and has one opening. Applications close on 15 November 2026.
This is a project-ready role rather than an introductory AI course. The employer specifically requires at least one end-to-end tabular machine-learning project and experience training a neural network. Applicants should be able to explain their practical work as well as demonstrate the Python, mathematics and model-evaluation foundations listed below.
Internship details
- Employer: DSights Consultancy Services Private Limited.
- Role: Machine Learning Engineer Intern.
- Mode: Work from home.
- Duration: Four months.
- Advertised stipend: ₹20,000–30,000 per month.
- Openings: One.
- Start window: 15 October–19 November 2026.
- Deadline: 15 November 2026.
What the intern will work on
The work uses customer, transaction, financial and operational datasets. The intern will build, evaluate, explain and integrate predictive models, with particular emphasis on tabular information. This is a useful distinction for applicants whose experience is mainly with chatbots or image generation: the primary assignment described here is not generic prompt writing.
Data preparation includes handling missing values, outliers, categorical variables and other quality issues. The intern will perform exploratory analysis and feature engineering, then build classification and regression models against clear baselines. The employer also wants comparisons between tree-based ensembles and neural networks for tabular prediction.
Evaluation and reproducibility are important parts of the role. The intern will design training, validation and test splits, prevent data leakage, choose suitable metrics, run cross-validation and investigate errors. Trained models should be packaged for repeatable inference rather than left as a one-off notebook result.
- Build reproducible Python pipelines for preparation, training and inference.
- Document experiments and explain model predictions.
- Compare approaches rather than reporting only a favourable result.
- Collaborate with developers on integrating models into applications.
Required technical skills
Strong Python skills are required, including pandas, NumPy and scikit-learn, together with working knowledge of SQL and Git. The employer also lists probability, statistics and linear algebra as necessary foundations. Applicants should understand regression, classification, decision trees, random forests and gradient boosting.
Hands-on experience with XGBoost, LightGBM or CatBoost is requested. For deep learning, the listing asks for understanding of neural networks, backpropagation, loss functions, optimisation and regularisation, plus practical experience with PyTorch or TensorFlow. Feature encoding, class imbalance, overfitting, cross-validation and evaluation are also explicitly named.
The project requirement is material: at least one end-to-end tabular ML project and experience training a neural network are required. Merely being interested in AI, or having watched a few tutorials without practical work, does not match the experience described in the listing.
Eligibility and useful advantages
Students or recent graduates in Statistics, Mathematics, Engineering, Computer Science, AI/ML or related quantitative fields may apply. The complete visible employer eligibility is not restricted only to Computer Science. Applicants must have relevant skills, be available for the full four months and be able to start within the stated window.
Women starting or restarting their careers are also welcome. The employer's Bangalore base is not an in-office instruction: this particular listing explicitly says work from home. No particular daily working hours or five-day week are promised in the visible terms, so confirm those details directly rather than assuming them.
Useful additional exposure includes TabNet, FT-Transformer, TabPFN or TabICL; SHAP; probability calibration; synthetic data; hyperparameter tuning; experiment tracking; and deployment. Interest in credit risk, churn, customer propensity, collections or forecasting is also mentioned. These are listed as good-to-have areas, not replacements for the required foundations and projects.
Stipend, perks and possible progression
The advertised pay range is ₹20,000–30,000 per month. The visible listing does not separate this into fixed and incentive components. Confirm the final offer amount, payment schedule, responsibilities and supervision in writing before accepting.
A certificate and letter of recommendation are listed. The source also mentions a potential pathway to a full-time Machine Learning Engineer role, but explicitly makes it subject to performance and business requirements. This is not a guaranteed permanent appointment, and no permanent salary figure is provided.
Preparing an application
As practical preparation, choose a project you can explain from data preparation through evaluation. Be ready to describe your baseline, split strategy, metric choice, leakage checks and error analysis, as well as your own contribution. A concise explanation of what did not work can be as useful as showing a final score. This is preparation guidance, not an additional employer assessment announced here.
Read the complete live opportunity and apply through Internshala by 15 November 2026. The original posting date is 16 October 2026, and applications were open when verified.
Read the complete DSights machine-learning internship and apply on Internshala.
