Machine Learning Engineer (Core Modelling)

Indpro AB; Bengaluru; On-site; INR 1,500,000 - 2,100,000; Full time

Job summary

Indpro AB is seeking a Machine Learning Engineer focused on core modelling to build, train, evaluate, and deploy robust ML models in production. You will work across time-series, probabilistic, and deep learning approaches, ensuring sound validation and leakage prevention while leveraging GenAI only when it adds genuine value.

The role emphasizes framing business problems, selecting appropriate models, and proving performance under real-world data challenges, with production-grade deployment and

Qualifications

• Genuine core-ML fundamentals explained from first principles.
• A real track record of models built and evaluated with architecture, data, metric, and before/after numbers.
• Fluency in Python stack and at least one deep learning framework for real training.
• Sound experimental discipline and honest, numbers-driven evaluation.

Responsibilities

• Own ML problems end-to-end from framing to production deployment.
• Build and optimise classical, probabilistic, time-series, and deep learning models.
• Design robust feature pipelines and prevent data leakage through proper validation.
• Select meaningful metrics and defend them with evidence.
• Deploy, monitor, and retrain models with drift detection.
• Leverage GenAI judiciously where it adds value to modelling, not as a substitute.

Skills

Core ML fundamentals

Model framing & evaluation

Python & scikit-learn

PyTorch or TensorFlow

SQL & data wrangling (Pandas)

Feature engineering

Cross-validation & leakage prevention

Production ML deployment

Education

Master's or PhD in CS/Math/Statistics

Tools

XGBoost

LightGBM

CatBoost

Pandas

SQL

Job description

Machine Learning Engineer · Core Modelling

Build models that matter. In production.

We build machine learning systems that make real decisions in production — forecasting, anomaly detection, ranking, classification, and probabilistic modelling. This is a core modelling role: understanding models deeply, not just orchestrating existing AI services.

Focus Core modelling — not GenAI orchestration.

About the role

This is a core machine learning role — not a prompt engineering or RAG engineering position. We use LLMs, retrieval, and GenAI where they add value, but our focus is building, training, evaluating, and deploying robust machine learning models. We want someone who understands modelling deeply rather than simply wiring together existing AI services.

If the most interesting thing you did last year was pick a loss function, catch a subtle leak that was inflating your metrics, or prove a simpler model beat a fancier one — we want to talk to you.

What you’ll do

Take an ambiguous business problem, frame it correctly, choose and train the right model, and prove it works under noisy data, distribution shift, class imbalance, and the constant threat of leakage.

• Own ML problems end-to-end — from problem framing to production deployment.
• Build and optimise classical, probabilistic, time‑series, and deep learning models based on the problem.
• Design robust feature‑engineering pipelines and prevent data leakage through sound validation.
• Select evaluation metrics that fit the problem, and defend them with evidence.
• Deploy, monitor, and improve production models — with retraining and drift monitoring.
• Use GenAI where it genuinely improves the solution — not as a substitute for modelling.
• Communicate modelling decisions clearly and shape the technical direction of our ML practice.

Tech you’ll use

Depth in the modern ML stack — used for real training and evaluation, not just inference.

Languages & data

Core ML

Tree ensembles

XGBoost, LightGBM, CatBoost

Deep learning

CNNs, LSTMs, Transformers

MLOps

What we’re looking for

Required fundamentals

We hire on evidence and first principles. You should be able to explain why a model overfits, what a loss function optimises, and why your validation scheme is sound for the data in front of you.

Specifics over slogans. “Implemented an ML model for various use cases” is not a track record. Bring the architecture, the data, the metric, and the before/after numbers.

• Genuine core-ML fundamentals. You can explain, from first principles: the bias–variance trade‑off, overfitting, what a loss function optimises, how regularization works, and why your validation scheme is sound.
• A real track record. Models you actually built and evaluated — with the architecture, the data, the metric, and the before/after numbers.
• Fluency in the modern ML stack. Python and scikit‑learn, plus at least one of PyTorch or TensorFlow used for real training. Solid SQL and data‑wrangling (pandas or PySpark) on large datasets.
• Sound experimental discipline. Appropriate cross‑validation, leakage prevention, metric selection, and the judgment to prefer a simpler model when it wins.
• The instinct to quantify — and honesty. You reach for numbers by default, know the difference between a self‑reported and an externally validated metric, and admit when a result doesn’t hold up.

Nice to have

Preferred, not required.

Probabilistic & Bayesian modelling

Time‑series beyond the standard toolkit — state‑space, hierarchical, or foundation models for forecasting. Architectures built or meaningfully modified from scratch, training‑dynamics debugging, or published research.

Genuine fine‑tuning experience

LoRA, PEFT, or full fine‑tunes with a clear eval story — and precisely what it accomplished over prompting.

Production MLOps maturity

CI/CD for models, drift detection (PSI or Evidently), automated retraining, canary releases, and observability.

Applied GenAI / RAG

Welcome as a complement to modelling depth — not a substitute for it.

Competition results, open‑source ML contributions, or peer‑reviewed publications.

What this role isn’t

Let’s save us both time.

These are real, valuable skills — they’re just not what this position is for. We hire for those on our Applied GenAI team, and we’re happy to redirect strong applicants there.

• Prompt engineering and LLM orchestration as the primary skill.
• Building RAG pipelines (chunking, embedding, vector search) without underlying modelling work.
• Agentic‑workflow wiring (LangChain, LangGraph, CrewAI) as the core of your portfolio.
• Integrating third‑party model APIs without training or evaluating models yourself.

Shape our ML practice.

You’ll help define how machine learning is practised across the company, influence architectural decisions, mentor engineers, and build production systems that solve meaningful business problems.