Data Operations Associate - ML Delivery

Location: Gurugram (Hybrid: 2 to 3 days on-site)

Experience: 1 to 3 years (Fresh graduates with strong fundamentals are welcome)

About Mechademy

Mechademy builds an enterprise AI platform for real-time monitoring, diagnostics, and predictive maintenance of industrial equipment. We are a venture-funded, Series A closed company experiencing 8090% YoY growth, with over 50 employees across New Delhi and Houston. Our platform serves 15+ major enterprise clients across oil & gas, power generation, and LNG, leveraging physics-informed machine learning and production AI to maintain a 12–18 month lead over the field.

Job Overview: The Opportunity

Most entry-level data roles leave you stuck behind a dashboard or inside a Jupyter notebook. This role puts you at the controls of a production ML operation.

You will join a high-impact operations function owning the delivery backbone of Mechademy’s AI platform. We run hundreds of predictive models across enterprise clients through a repeatable lifecycle: creating experiments, training models, validating them against fresh operational data, deploying them, and monitoring accuracy over time.

If you care more about things working reliably than looking clever, and want to own the machine learning lifecycle in production, this is for you.

Key Responsibilities

1. Model Delivery & Lifecycle (45%)

- Take predictive models from request to live client equipment and keep them healthy.
- Run internal tooling to generate, validate, and deploy predictive models following strict runbooks.
- Process and validate incoming sensor data using structured checklists (spotting missing values, frozen readings, type/unit mismatches, and outliers).
- Verify model behavior post-deployment and track lifecycle status across clients (live, in-progress, blocked).




- Monitor deployed models over time to flag accuracy drift and degradation.

1. Automation & Operational Reliability (30%)

- Maintain internal tooling, runbooks, SLA dashboards, and operational trackers as the single source of truth.
- Reconcile internal records against the live platform to ensure zero data discrepancies.
- Transition from executing manual steps to overseeing automation quality as tooling improves.

1. Documentation & ML Process Improvement (25%)

- Document operational procedures clearly so new team members can ramp up seamlessly.
- Research portfolio growth applications (e.g., comparing regression vs. anomaly detection, physics-based multivariate models vs. data-driven autoencoders) and document trade-offs.
- Propose and test process/monitoring enhancements to prevent production issues.

What Success Looks Like
- First 30 Days: Successfully completed your first supervised model deployment end-to-end; navigate tooling and runbooks with guidance.
- First 60 Days: Independently running validations and deployments, applying readiness criteria without hand-holding, and maintaining a near-zero error rate.
- First 90 Days: Full autonomy over a segment of the model pipeline, contributing to automation improvements and trusted completely on model health sign-offs.

Requirements & Qualifications (Must-Haves)

- Education: B.Tech / B.E. / B.S. in Computer Science, IT, Electronics, Mathematics, Statistics,



or related technical fields.
- Experience: 1–3 years; fresh graduates with strong fundamentals and relevant academic or internship experience are welcome.
- Python & Scripting Comfort: Ability to read/run scripts, use the command line, edit config files, and automate small tasks.
- Data Literacy: Strong grasp of tabular data, anomaly/outlier detection, and the ability to interpret time-series sensor data (drifts, flatlines, gaps, step changes).
- ML Knowledge & Judgment: Familiarity with data/ML tooling (pandas, experiment tracking, error metrics, notebooks). Capable of critically evaluating validation metrics rather than taking positive-looking numbers at face value.
- Rigor & Ownership: Process-driven mindset, meticulous attention to detail under volume, and clear written communication skills.

Strong Signals & Right Mindset
- Reliability over Cleverness: You prefer doing the boring thing correctly every time over inconsistent heroics.
- Signal over Noise: When flagging issues, you provide precise diagnostic context (what is wrong, where it is, and what should happen next).
- Suspicious of Positive News: You treat validation metrics as claims to be verified rather than absolute conclusions.
- Cloud/Domain Familiarity: Exposure to AWS/S3, APIs, or industrial/energy operations is a strong plus.

What We Offer

- Production AI Experience: Hands-on exposure to enterprise-scale AI infrastructure from day one.
- Real Impact: Your discipline directly determines client trust and operational success in critical industrial environments.
- Rapid Growth: A meritocratic environment where scope expands alongside your reliability.
- Hybrid Flexibility: 2–3 days on-site in Gurugram, with competitive compensation matching industry standards. .

📌 Hiring Data Operation Associate (Delhi)
🏢 Mechademy
📍 Delhi