Ecolab
Website:
ecolab.com
Job details:
Connected Assets
600,000+ across 40+ industries globally
Digital Products Powered
13 (WQ IQ, CIP IQ, Aqua IQ, Dish IQ, Pest Intelligence, Kitchen IQ, HX IQ, and more)
Digital Revenue Enabled
$370M+ annually
Data Points Processed
250+ billion per year from industrial IoT controllers
Team Size
150+ domain experts across 6 global locations
Global Locations
Pune (India), Naperville (USA), Jeddah (Saudi Arabia), and 3 additional sites
Alarm Automation Rate
88% and growing through AI-driven transformation
TVD Generated
$22M in 2025, on track for $100M+ in 2026
EGIC 2.0 Transformation Context
EGIC Is Undergoing a Fundamental Transformation From Reactive Operations (EGIC 1.0) To An Autonomous Intelligence Engine (EGIC 2.0) Through Two Strategic Programs
- AIM (AI Integration & Modernization): Embedding AI across all EGIC workflows — automating reactive alarm triage, building predictive intelligence, and liberating 25–30 FTE-equivalent capacity through agentic automation.
- APEX (Adoption, Proactive Engagement & eXpansion): Driving digital adoption, converting hardware-only customers to digital subscribers, and scaling TVD from $22M to $100M+ through proactive customer engagement.
This role is central to both programs — the candidate will lead the AI team that powers AIM and provides the intelligence backbone for APEX.
3.1 The Opportunity
We are looking for a Sr AI Engineer — a rare hybrid professional who combines deep business process understanding with advanced AI/Data Science technical expertise and proven people leadership. This individual will lead a team of 5–6 data scientists and AI engineers, driving the daily AI operations, project delivery, model deployment, adoption enablement, and production triage across EGIC's entire digital product portfolio.
The ideal candidate has walked a deliberate career path: starting with understanding how business processes work (BPA), then building the AI that transforms those processes (Data Science), and now leading the team that operates, scales, and governs that AI at enterprise grade (Leadership). This combination ensures the candidate doesn't just build technically impressive models, but builds the RIGHT models that solve REAL operational problems and deliver MEASURABLE business outcomes.
3.2 What Makes This Role Unique
- You will lead AI operations for one of the largest industrial IoT intelligence centers in the world — 600K+ connected assets, 250B+ data points/year
- Your models will directly impact water conservation, energy efficiency, and sustainability outcomes for customers across 40+ industries globally
- You will have end-to-end ownership: from identifying the business problem (BPA lens) → designing the AI solution (DS lens) → deploying and scaling it (Ops lens) → measuring its impact (Leadership lens)
- You will build and shape a team from a position of influence — defining the culture, technical standards, and the operational playbook
- Direct visibility to EGIC Director and senior leadership — your work directly drives strategic decisions and commercial outcomes ($100M+ TVD target)
- Cutting-edge technology: Databricks, Azure, LLMs, agentic AI, RAG pipelines, edge computing, real-time IoT streaming
3.3 This Role Is NOT
- NOT a pure research/academic role — we need production engineers, not just experimenters
- NOT a people-only management role — you must be technically hands-on and review architecture decisions
- NOT a single-project role — you will manage a portfolio of 3–5 concurrent AI projects across different digital products
- NOT isolated from the business — you will regularly interact with Operations, Product, Engineering, Field, and Customer-facing teams
- Key Responsibilities
4A. Business Process Analysis & Operations Leadership
Your BPA foundation differentiates this role from a typical Data Science lead. You will use process analysis expertise to ensure every AI solution is grounded in a deep understanding of the operational workflow it serves. AI without process understanding leads to technically impressive but operationally useless solutions.
i. End-to-End Process Mapping & Analysis
- Map all EGIC operational workflows across 13 digital products — from data ingestion → alarm generation → triage → insight creation → customer delivery → value capture (TVD)
- Identify human touchpoints, decision nodes, escalation paths, and handoff points in each workflow
- Quantify time spent, error rates, and throughput at each process step to identify automation ROI
- Use BPMN (Business Process Modeling Notation) or equivalent tools to create standardized, version-controlled process documentation
ii. Automation Opportunity Identification
- Systematically identify and prioritize automation candidates using a structured scoring framework (Impact × Feasibility × Urgency)
- Distinguish between rule-based automation (RPA), ML-based automation (predictive models), and agentic automation (LLM-powered agents) — recommending the right approach for each use case
- Build and maintain an "Automation Opportunity Backlog" — a prioritized pipeline of process improvements the AI team will deliver against
- Conduct JTBD (Jobs To Be Done) analysis for each manual workflow to understand the root purpose before automating
iii. SOP Design & Governance
- Design and document Standard Operating Procedures (SOPs) for all AI-augmented workflows — covering normal operations, exception handling, escalation protocols, and fallback procedures
- Ensure SOPs are practical, maintainable, and actually followed by operations teams — not just shelfware
- Establish a quarterly SOP review cadence to incorporate lessons learned from production incidents and process changes
iv. Operational KPI Framework
Define, instrument, and track operational KPIs across all AI-augmented workflows:
- SLA adherence rate (% of alarms/insights delivered within defined timeframes)
- Automation rate (% of alarms handled without human intervention)
- Triage turnaround time (mean time from alarm to resolution)
- FTE liberation (hours/FTE equivalent freed through automation)
- Model uptime (% availability of deployed AI models)
- Inference latency (p50, p95, p99 response times for real-time models)
- False positive/negative rates for alarm classification models
- Build automated KPI dashboards (Power BI / Databricks) and publish weekly operational reports to leadership
v. Cross-Functional Process Alignment
- Facilitate process alignment workshops between Operations, Engineering, Product, and Field teams
- Ensure AI outputs integrate seamlessly into existing operational tools (ServiceMax, ECOLAB3D, Salesforce)
- Act as the "process conscience" of the AI team — ensuring solutions are designed for the real operational context
vi. AIM Program Contribution
- Contribute to EGIC's AIM (AI Integration & Modernization) program by identifying, chartering, and executing rationalization and automation projects
- Lead process discovery sessions for new AIM workstreams — understanding current-state workflows before designing AI solutions
- Track and report AIM impact metrics: FTE liberated, hours saved, error reduction, customer impact
4B. AI / Data Science Technical Leadership
This is where your Data Science expertise drives technical vision. You will own the architectural integrity of all AI solutions, ensuring they are not just accurate but production-grade, scalable, maintainable, and cost-efficient. You must be hands-on enough to review code and architecture — and strategic enough to choose the right battles.
i. AI/ML Model Design & Development
Lead the design, development, and deployment of ML/AI models for EGIC's core industrial IoT use cases:
- Alarm Auto-Classification: Multi-class models that classify, prioritize, and route alarms across Water, Institutional, and Pest products
- Predictive Maintenance: Time-series models predicting asset failures (pumps, controllers, sensors) before they impact operations
- Anomaly Detection: Unsupervised models detecting novel patterns in industrial telemetry data (water quality, chemical dosing, temperature, conductivity)
- NLP-Based Report Generation: LLM-powered systems auto-generating customer-facing reports, business reviews, and insight summaries
- Agentic Automation: Multi-agent systems orchestrating complex operational workflows (alarm → analysis → recommendation → delivery) with minimal human intervention
- TVD Discovery: AI models scanning connected assets to identify, quantify, and dollarize value opportunities for customers
ii. Architecture Review & Governance
- Review and approve ALL model architectures proposed by team members before development begins
- Ensure correct architecture selection: batch vs. real-time inference, cloud vs. edge deployment, rule-based vs. ML vs. LLM-based approaches, monolithic vs. microservices model serving
- Maintain an Architecture Decision Record (ADR) documenting rationale behind key technical choices
- Conduct monthly "Architecture Review" sessions where the team presents and defends design decisions
iii. GenAI, LLM & Agentic AI
- Lead evaluation and integration of GenAI technologies: RAG pipelines, LLM-powered conversational agents, agentic frameworks (LangChain, LangGraph, CrewAI)
- Establish prompt engineering standards, prompt versioning, and LLM evaluation frameworks (hallucination rate, relevance, groundedness, safety)
- Ensure guardrails and governance for GenAI: compliance with Ecolab AI ethics, data privacy, and IP protections
iv. Data Engineering Collaboration
- Work with Data Engineers to ensure clean, reliable, timely data pipelines feeding ML models
- Define data contracts between data engineering and data science teams — specifying schema, SLAs, and quality expectations
- Oversee real-time streaming pipelines (Kafka / Azure Event Hubs) for time-sensitive use cases
v. Innovation & Emerging Technology
- Stay current with AI/ML research; evaluate emerging technologies (multimodal AI, Small Language Models, Graph Neural Networks, Reinforcement Learning, Federated Learning)
- Run structured PoC → Pilot → Production cycles for promising technologies
- Present quarterly "Technology Radar" updates to leadership highlighting emerging capabilities and EGIC relevance
4C. Team Leadership & People Management
You will build, lead, and develop a team of 5–6 data scientists and AI engineers. This is not just task allocation — it's about creating a high-performing team culture where technical excellence, ownership, and continuous learning are the norm. You are expected to be a player-coach.
i. Team Building & Culture
- Build a culture centered on: technical rigor, ownership, transparency, continuous learning, and collaborative problem-solving
- Establish team rituals: daily standups, weekly technical deep-dives, monthly innovation hours, quarterly retrospectives
- Create a psychologically safe environment for experimentation, fast failure, and improvement
- Foster knowledge sharing through internal tech talks, code review sessions, and documentation standards
ii. People Management
- Conduct weekly/bi-weekly 1:1s — covering project progress, blockers, career aspirations, and wellbeing
- Lead quarterly performance reviews aligned with Ecolab's VSEM framework
- Create Individual Development Plans (IDPs) for each team member — mapping skill gaps, training, and career progression
- Handle team dynamics, conflict resolution, and performance improvement with empathy and professionalism
- Participate in hiring: interviews, candidate evaluation, and onboarding of new team members
iii. Project & Delivery Management
- Manage team capacity across 3–5 concurrent projects — balancing business priority, member growth, and delivery timelines
- Run Agile/Scrum sprint cycles (2-week sprints) with planning, estimation, execution, and retrospective ceremonies
- Maintain project portfolio dashboard: status, milestones, risks, dependencies, and resource allocation
- Ensure every project has a clear charter: problem statement, success criteria, data requirements, architecture, timeline, and ownership
iv. Upskilling & Professional Development
- Drive structured upskilling: Databricks certifications, Azure certifications, LLM/GenAI training, MLOps best practices, soft skills development
- Allocate 10% of sprint capacity for self-directed learning and experimentation
- Encourage conference presentations, blog posts, and internal knowledge sharing
4D. Deployment, Adoption & Production Triage
A model that works in a notebook but fails in production is not a success. This section ensures every AI solution survives — and thrives — in the real operational environment.
i. Production Deployment & Release Management
Every Deployment Follows a Standardized Release Checklist
- Code review completed and approved
- Unit tests and integration tests passing
- Performance benchmarks met (accuracy, latency, throughput)
- Data quality checks validated
- Rollback plan documented and tested
- Monitoring and alerting configured
- Stakeholder sign-off obtained
- Implement blue-green or canary deployment strategies for high-risk model updates
ii. Production Monitoring & Health
- Build comprehensive monitoring dashboards: model performance, data quality, inference latency, business impact, infrastructure utilization
- Configure automated alerting for performance degradation, data drift, concept drift, and infrastructure issues
- Publish weekly "Production Health Report" summarizing model status, incidents, and actions
iii. Incident Triage & Resolution
Establish And Manage a Structured Triage Process
- Severity Classification: P1 (critical — customer impact, 30-min response), P2 (high — 2-hr response), P3 (medium — 1 business day), P4 (low — 1 week)
- Mandatory Root Cause Analysis (RCA) for all P1 and P2 incidents with post-mortem documentation
- Shared on-call rotation across 5–6 team members for after-hours P1/P2 incidents
- Maintain incident knowledge base documenting past incidents, root causes, and resolutions
iv. Adoption Enablement
- Drive adoption of AI solutions across operations — ensure teams USE and TRUST AI outputs
- Conduct training sessions and workshops for operations teams on interpreting and acting on AI insights
- Create user guides, quick-reference cards, and FAQ documents for each deployed solution
- Establish feedback loops: collect user feedback on accuracy / usability / trust; feed back into model improvement
- Track adoption metrics: % of AI insights acted upon, user satisfaction, time-to-action
v. Documentation & Knowledge Management
- Maintain comprehensive documentation: model cards, API docs, data dictionaries, runbooks, architecture diagrams
- Ensure documentation is current, searchable, and accessible to all relevant stakeholders
- Required Qualifications
5.1 Education
- B.E. / B.Tech in Computer Science, Data Science, Information Technology, Chemical Engineering, or related engineering discipline — Required
- M.S. / M.Tech in Data Science, AI/ML, Statistics, Applied Mathematics, or Computational Engineering — Preferred
- MBA or PG Diploma in Business Analytics / Operations Management — Strong plus for the BPA dimension
Valued Certifications
- Databricks Certified Data Scientist / ML Professional
- Azure Data Scientist Associate (DP-100) / Azure AI Engineer Associate (AI-102)
- Lean Six Sigma Green/Black Belt
- PMP / PMI-ACP / Certified Scrum Master
5.2 Experience Breakdown
The ideal candidate has a blended career trajectory spanning three distinct capability areas:
Capability Area
Years
What We Expect
Demonstrated By
Business Process Analysis / Process Engineering
3–4 yrs
Workflow design, process mapping (BPMN), SOP creation, operational KPIs, stakeholder interviews, gap analysis, Lean/Six Sigma, cross-functional facilitation
Process documentation portfolio, efficiency metrics (% time saved, error reduction), SOP samples, process audit reports
Data Science / AI Engineering / ML Engineering
3–4 yrs
Hands-on ML model building & production deployment, ML lifecycle management, cloud architecture (Azure/Databricks), GenAI/LLM development, time-series analysis, anomaly detection, NLP, MLOps
Production models deployed, GitHub/code portfolio, MLOps pipeline examples, architecture decision records, model performance dashboards
Team Leadership / People Management
1–2 yrs
Leading 3+ technical professionals, performance reviews, sprint/project planning, mentoring, stakeholder management, delivery ownership, hiring participation
Team size managed, projects delivered, team member growth, stakeholder testimonials, delivery track record
5.3 Technical Skills — Must Have
Category
Skill
Level
Context
Programming
Python
Expert
OOP, Pandas, NumPy, scikit-learn, XGBoost, production coding, unit testing
Programming
SQL & PySpark
Expert
Complex queries, distributed processing, Delta Lake operations
ML / AI
Classical ML
Expert
Classification, regression, clustering, ensemble methods, time-series
ML / AI
Deep Learning
Advanced
TensorFlow or PyTorch for production deployments
ML / AI
Anomaly Detection
Advanced
Isolation Forest, autoencoders, statistical methods — core EGIC use case
ML / AI
NLP
Advanced
Text classification, NER, summarization, sentiment analysis
MLOps
Databricks
Expert
MLflow, Unity Catalog, Feature Store, Model Serving, Delta Live Tables
MLOps
CI/CD for ML
Advanced
Azure DevOps pipelines, automated testing, model validation gates
Cloud
Azure
Advanced
ADF, Databricks, Azure ML, Blob Storage, Key Vault, Azure DevOps
GenAI
LLM Frameworks
Advanced
LangChain, LangGraph, CrewAI — agentic applications
GenAI
RAG Pipelines
Advanced
Vector DBs, embedding models, retrieval strategies, chunking
GenAI
Prompt Engineering
Advanced
System prompts, few-shot, chain-of-thought, prompt versioning
Visualization
Power BI / Streamlit
Advanced
Production dashboards, DAX, data modeling
Version Control
Git & Azure DevOps
Advanced
Branching strategies, PRs, code reviews
5.4 Technical Skills — Nice to Have
- Edge AI / IoT deployment (Raspberry Pi, NVIDIA Jetson, TensorFlow Lite, ONNX Runtime)
- Computer Vision for industrial applications (defect detection, gauge reading, visual inspection)
- Databricks Apps, Genie Spaces, AI/BI Dashboards
- Microsoft Copilot Studio, Power Platform, Power Automate
- Snowflake (EGIC uses both Databricks and Snowflake)
- Kubernetes / Docker for containerized ML model deployments
- Agent-to-Agent (A2A) and Model Context Protocol (MCP) architectures
5.5 Domain Knowledge — Critical
This role operates at the intersection of AI and industrial operations. Domain knowledge is not optional — it is what separates a good data scientist from one who can deliver real impact at EGIC.
- Industrial Water Treatment: Cooling tower chemistry, boiler water management, RO/membrane systems, wastewater treatment, chemical dosing, corrosion/scale control. Understanding why parameters like conductivity, pH, ORP, turbidity, and cycles of concentration matter operationally.
- IoT / Industrial Telemetry: Real-time sensor data from 3D TRASAR, AIO controllers, OMNI systems. Alarm management principles (ISA-18.2), sensor calibration and drift, SCADA-level concepts. Understanding data flow from field devices through gateways to cloud.
- 24/7 Operations & Service Delivery: SLA-driven operations, shift-based monitoring, escalation protocols, incident management (ITIL-aligned). Understanding the operational tempo and pressure of a global intelligence center.
- Industry Preference: Experience at Nalco Water, Ecolab, Veolia, Suez, Xylem, Kemira, BASF, Dow, Honeywell Process Solutions, Emerson, or Schneider Electric is highly preferred.
- Platform Familiarity: ECOLAB3D, ServiceMax, Salesforce, or equivalent industrial IoT and field service platforms.
5.6 Soft Skills & Competencies
- Leadership & Influence: Lead without authority, influence cross-functional stakeholders, drive alignment. Comfortable presenting to Director-level and above.
- Communication Excellence: Explain complex AI/ML concepts to non-technical audiences. Strong written communication for docs, reports, and executive summaries.
- Process-Oriented Mindset: Natural inclination to understand, document, and improve processes. Bias for systematic approaches over ad-hoc solutions.
- Bias for Action: Make decisions with imperfect information. Preference for "good enough now" over "perfect later" — especially in production triage.
- Problem-Solving Under Pressure: Calm, structured, time-sensitive decision-making in production environments.
- Cross-Functional Fluency: Fluent in technology, operations, and business languages. Can translate between all three seamlessly.
- Growth Mindset: Genuine curiosity about new technologies, willingness to learn, adapt, and teach others.
Nuestro compromiso con una cultura de inclusión y pertenencia
Ecolab está comprometido con el trato justo e igualitario de todas las personas colaboradoras y postulantes, y con la promoción de los principios de igualdad de oportunidades en el empleo. Reclutaremos, contrataremos, promoveremos, transferiremos y brindaremos oportunidades de desarrollo con base en las calificaciones individuales y el desempeño laboral, en todos los aspectos relacionados con el empleo, la compensación, los beneficios, las condiciones laborales y las oportunidades de crecimiento. Ecolab no discriminará a ninguna persona colaboradora ni postulante por motivos de raza, religión, color, credo, nacionalidad, estado de ciudadanía, sexo, orientación sexual, identidad y expresión de género, información genética, estado civil, edad o discapacidad.
Click on Apply to know more.