Senior Machine Learning Engineer

VuNet Systems; Bengaluru; On-site; INR 4,200,000 - 7,000,000; Full time

Job summary

VuNet Systems in Bengaluru seeks a Senior Machine Learning Engineer to architect and productionize ML models for observability across time-series data, ensuring low-latency inference and scalable deployment. You will collaborate with platform engineering, SRE and product teams to translate first principles into robust ML pipelines, focusing on anomaly detection, forecasting and RCA.

Candidates should have strong Python software engineering, statistics, and experience deploying ML in distributed

Qualifications

• Strong software engineering skills in Python and production-grade code.
• Solid grounding in probability, statistics, statistical inference and ML.
• Good understanding of time-series analysis: trends, seasonality, forecasting, change detection.
• Ability to reason about algorithms beyond library APIs: assumptions, trade-offs, computational complexity.
• Experience building or adapting ML/statistical algorithms to production, not just integration.
• Experience taking algorithms from experiments to reliable production systems.
• Strong debugging and analytical problem-solving skills.
• Understanding of distributed systems and high-volume data processing.
• Ability to work with imperfect, noisy datasets.
• High agency: identify problems, prototype solutions, validate with data.

Responsibilities

• Enhance and productionize VuNet's ML/MLOps platform for building, orchestrating, deploying, validating and monitoring ML workloads.
• Work with Temporal-based workflows and VuNet's ML frameworks for model execution, scheduling and lifecycle management.
• Build mechanisms for model versioning, experimentation, benchmarking, validation, safe rollout and performance monitoring.
• Engineer ML capabilities for continuous production workloads across thousands to tens of thousands of time series and monitored entities.
• Design for high-throughput streaming data, distributed computation, low-latency inference, horizontal scalability and resource efficiency.
• Develop adaptive anomaly-detection algorithms across infrastructure, application, transaction and business time-series data.
• Build techniques that account for seasonality, trends, changing baselines, noise and workload patterns.
• Explore change-point detection and distribution-shift techniques to identify meaningful behavioural changes.
• Explore multivariate approaches where anomalies emerge from relationships among signals.

Skills

Python

Probability

Statistics

Time-series

Distributed systems

Debugging

High agency

Job description

VuNet is a pioneer in Business Journey Observability, leveraging Big Data and Machine Learning to transform digital experiences across the financial services. Our deep-tech platform provides end-to-end visibility into customer journeys — empowering proactive issue resolution, operational resilience, and superior user satisfaction.

If you’ve ever used instant payment systems like UPI, chances are you’ve already experienced the power of our platform — we monitor over 28 billion digital transactions monthly (that’s equal to watching 3 years of tik-tok videos), touching 400 million users with leading banks and financial institutions.

VuNet is Series B funded, part of NASSCOM DeepTech Club, awarded NASSCOM’s AI Gamechanger, recognized in Forbes DGEMS 200 and by several global analysts including Gartner, Omdia.

We’re building a new category of observability purpose-built for complex digital journeys — across payments, lending, core banking and more — already powering some of the largest banks in India and MEA.

Your Role: Senior Machine Learning Engineer

We are looking for a Senior Machine Learning Engineer who combines strong mathematical and statistical foundations with solid software engineering.

This is not primarily a GenAI, prompt-engineering or AI-API integration role.

You will build the underlying intelligence used by our observability platform: statistical models, machine-learning algorithms and analytical techniques that operate continuously on large volumes of time-series and operational data.

We are looking for someone who can reason from first principles, formulate a problem mathematically, evaluate alternative approaches, understand their assumptions and failure modes, implement the solution and take it all the way into reliable production.

You will work closely with platform engineering, SRE, product and domain teams, but will be expected to independently identify opportunities where better algorithms can materially improve the product.

ML Platform & Production Engineering

• Enhance and productionize VuNet's ML/MLOps platform for building, orchestrating, deploying, validating and monitoring ML workloads.
• Work with Temporal-based workflows and VuNet's ML frameworks for model execution, scheduling and lifecycle management.
• Build mechanisms for model versioning, experimentation, benchmarking, validation, safe rollout and performance monitoring.
• Engineer ML capabilities for continuous production workloads across thousands to tens of thousands of time series and monitored entities.
• Design for high-throughput streaming data, distributed computation, low-latency inference, horizontal scalability and resource efficiency.
• Develop adaptive anomaly-detection algorithms across infrastructure, application, transaction and business time-series data.
• Build techniques that automatically account for seasonality, trends, changing baselines, noise and workload patterns.
• Explore change-point detection and distribution-shift techniques to identify meaningful behavioural changes.
• Explore multivariate approaches where anomalies emerge from relationships among signals rather than from individual metrics.

Forecasting & Predictive Operations

• Develop forecasting algorithms for capacity planning, workload forecasting, resource exhaustion and proactive operations.
• Build models capable of handling multiple seasonalities, incomplete/noisy telemetry and changing operational behaviour.
• Quantify prediction uncertainty and identify leading indicators that provide early warning of degradation or saturation.

Root Cause, Dependency and Incident Intelligence

• Develop analytical and ML techniques that distinguish probable root causes from downstream symptoms.
• Use service topology, dependencies, temporal relationships, behavioural correlation and statistical evidence for RCA.
• Correlate alerts, anomalies and operational events into meaningful incidents using time, topology, entity relationships and behavioural similarity.
• Develop techniques for event clustering, incident evolution, impact analysis, blast-radius identification and probable-cause ranking.
• Explore causal inference and dependency-aware approaches where they provide measurable value.

Algorithm Validation, Simulation & Quality

• Build systematic benchmarking and validation frameworks for anomaly detection, forecasting, correlation and RCA algorithms.
• Define evaluation measures including precision, recall, false-positive rate, detection delay, stability, forecasting error and operational usefulness.
• Evaluate algorithms under seasonality, concept drift, noisy or missing telemetry, cold-start conditions and changing workloads.
• Develop simulation and synthetic-data frameworks for realistic metrics, anomalies, incidents, failures and dependency scenarios.
• Validate algorithms against labelled datasets, historical production data and controlled or simulated scenarios to prevent regressions.

What You Bring

• Strong software engineering skills, particularly in Python.
• Strong grounding in probability, statistics, statistical inference and machine learning.
• Good understanding of time-series analysis including trends, seasonality, decomposition, forecasting and change detection.
• Ability to reason about algorithms beyond library APIs: assumptions, trade-offs, computational complexity, failure modes and appropriate evaluation methods.
• Experience building or materially adapting ML/statistical algorithms rather than only integrating pre-built AI services.
• Experience taking algorithms from experimentation into reliable production systems.
• Strong debugging and analytical problem-solving skills.
• Understanding of distributed systems and high-volume data processing.
• Ability to work with imperfect, noisy and evolving real-world datasets.
• High agency: identifies meaningful problems, forms hypotheses, prototypes solutions, validates them against data and drives successful approaches into production without waiting for detailed task definitions.

Good to Have Skills

• Go experience for production services or high-performance components.
• Experience with Apache Flink or similar streaming/data-processing frameworks.
• Experience working with observability, monitoring, SRE, telemetry or AIOps systems.
• Familiarity with metrics, logs, traces, service topology and operational event data.
• Experience with Kafka or other streaming platforms.
• Exposure to techniques such as Bayesian methods, clustering, dimensionality reduction, probabilistic models, causal inference or optimization.
• Experience building large-scale simulation, benchmarking or synthetic-data frameworks.
• Experience working on systems where false positives, model drift, latency and explainability have direct operational consequences.