Nagaraj arthem

I am a highly ambitious and diligent ML Data Engineer, driven by a logical and pragmatic approach to problem-solving. My passion lies in achieving mastery and bringing meticulous organization to complex data challenges, particularly within AI Observability and Generative AI, ensuring robust and efficient systems.

Experience

ML Data Engineer (Data Science & Machine Learning) | AI Observability & GenAI

Amazon · Jun 2024 – Present

Not yet confirmed

• ◗ Support the computer vision models behind robotic picking and packing across Amazon fulfillment centers → owning the data science and observability layer that keeps those models trustworthy after deployment, not just at launch.
• ↳ Explored 10K+ production inference/telemetry records across 4 ETL workflows using Python, Pandas, and Advanced SQL to find where prediction errors clustered by class, zone, and time
• ↳ Engineered 25+ features (confidence trends, class error-rate, conveyor congestion index) from model outputs and telemetry, served through a Feast feature store for train/serve consistency
• ↳ Trained and evaluated 6 classifiers (Logistic Regression through XGBoost) using cross-validation, precision/recall/F1, and ROC-AUC → improving one model's F1 from 63% to 91% and cutting error rate from 37% to 4.1% in 48 hours
• ↳ Logged experiments in MLflow → gated deployments through SageMaker Model Registry
• ↳ Built PSI/KS-test drift detection (input drift, output drift, production behavior) → exported as Prometheus metrics
• ↳ Designed 4 Grafana dashboards across 12+ model-health metrics → tuning alerts to cut false alarms → reducing mean-time-to-detect degradation from weeks to same-day
• ↳ Built a LangChain RAG copilot over 1K+ monitoring records: an LLM writes SQL grounded on an approved schema → a guardrail validates it pre-execution → results reconciled against a trusted figure → 86/100 test questions produced valid SQL, 81 reconciled correctly
• ↳ Fine-tuned a small LLM with LoRA/PEFT to summarize alert logs into evidence-checked root causes → cutting a 15-minute triage to a 20-second first draft (46/50 summaries fully evidence-grounded)
• ↳ Rebuilt a slow concept-drift SQL report with covering indexes and partition filters → cutting runtime from 8+ minutes to under 15 seconds
• ↳ Containerized jobs with Docker → wrote pytest coverage → shipped via GitHub Actions CI/CD in Agile sprints tracked in Jira

ML Data Associate | Business Intelligence

Amazon · Jun 2024 – Present

Not yet confirmed

• Enabled continuous improvement of Amazon Robotics models through MLOps, ETL Pipelines, data quality monitoring, and ML retraining support.
• Analyzed ML systems data using SQL and Tableau, transforming data into insights that improved model performance by 20%.
• Collaborated with cross-functional teams to optimize AWS MLOps pipelines and data workflows.
• Developed Tableau BI dashboards tracking annotation KPIs, increasing executive reporting efficiency by 45%.
• Conducted SQL and Python-based data quality assessments, identifying discrepancies and reducing data errors by 30%.
• Governed large-scale training datasets through ETL, data validation, and quality checks for ML model retraining.
• Presented actionable insights to stakeholders, translating data findings into risk management and business recommendations.

Data Analyst | SQL | Excel | Financial Forecasting |

B2B Logistics Pty Ltd · Mar 2023 – Jun 2024

Not yet confirmed

• ➜ Reviewed 50K+ shipment records using SQL and Python → surfacing delivery trends, delays, and recurring process gaps for operations stakeholders
• ➜ Derived 15+ analytical features from shipment and delivery data → supporting performance analysis across key logistics metrics
• ➜ Scheduled 6 recurring reports with Python and SQL → saving roughly 8 hours of manual data preparation per week
• ➜ Validated data accuracy feeding ML risk models → ensuring high-quality, client-ready deliverables
• ➜ Used Git branches and pull requests on every change → keeping a clean, auditable commit history

B

Data Analyst

B2B Logistics · Mar 2023 – Jun 2024

Not yet confirmed

• Assisted in maintaining SQL databases, improving data access speed by 25%.
• Used Excel and pivot tables for data analysis, supporting financial forecasting models.
• Supported senior analysts on finance projects, applying analytical and strategic thinking to solve business problems.
• Validated data accuracy on projects, ensuring high-quality deliverables and client-ready outputs.

Software Intern | SQL | Excel | Technical Reporting | GAOTek Inc

GAOTek Inc. · Sep 2022 – Jan 2023

Not yet confirmed

• Stack: SQL (Structured Query Language) | Excel | Technical Documentation
• -> Compiled 5+ weekly datasets in Excel for cross-team technical reports, keeping distributed teams aligned on shared metrics.
• -> Reduced manual lookup time nearly 20% by implementing basic SQL queries in place of manual spreadsheet searches.
• -> Documented 10+ process findings, contributing to a growing internal knowledge base and continuous learning initiatives.
• Tech stack: SQL | Excel | Technical Reporting | Process Documentation

G

Software Intern

GAO Tek Inc. · Sep 2022 – Jan 2023

Not yet confirmed

• Compiled 5+ weekly datasets in Excel for cross-team data reports.
• Reduced manual lookup time nearly 20% by implementing SQL basics.
• Documented 10+ process findings, contributing to a growing knowledge base and continuous learning initiatives.

What drives their work

The Architect

You build robust data and ML systems from the ground up.

Nagaraj's superpower

You excel at identifying data inconsistencies and engineering precise solutions that significantly enhance system performance and reliability.

How Nagaraj works

You thrive in teams focused on building and optimizing robust, data-driven systems with a clear impact.