About
Senior Data Scientist with 9+ years of experience delivering impactful machine learning solutions across industrial and SaaS domains. Expert in architecting and operationalizing predictive systems with ensemble modeling and production ML practices, including A/B testing-oriented optimization using business KPIs. Proven track record of deploying scalable inference services on AWS with Docker and Kubernetes, integrating explainability via SHAP into interactive Tableau dashboards, and improving outcomes such as 46% Closed-Won ARR and 15% downtime reduction. Seeking to contribute to Sonicgrid Technologies as a Machine Learning Engineer - Feed & Ranking Systems by building and optimizing ranking models driven by user engagement and measurable performance.
Skills & Expertise (28)
Work Experience
Data Scientist
Docusign (Randstad Payroll)
Jan 2025 - Present
Architected an end-to-end lead scoring pipeline on Amazon SageMaker, utilising SageMaker Pipelines for automated model retraining and driving a 46% increase in Closed-Won ARR. Engineered a prescriptive decision-optimisation layer using Gurobi downstream of SageMaker inference, transforming raw lead scores into optimal rep assignments under strict capacity and regional constraints to maximise ARR capture. Engineered production-grade codebases managed via Git/GitHub, implementing automated CI/CD workflows that reduced model deployment lead time by 40%. Containerised model inference services using Docker, ensuring environment parity and reducing “it works on my machine” deployment errors by 100%. Deployed high-throughput micro-services on Kubernetes, optimising resource orchestration to maintain sub-100ms inference latency during peak traffic. Built a SHAP-based interpretability framework integrated into interactive Tableau dashboards, cutting management review time by 80% and providing clear, actionable drivers of lead quality.
Data Scientist
ALSTOM (Randstad Payroll)
Apr 2021 - Sep 2023
Architected an end-to-end predictive maintenance pipeline on AWS for rolling stock brake systems, integrating real-time telemetry (TCMS) with wayside data to transition from reactive to condition-based maintenance and reduce unscheduled downtime by 15%. Engineered a hybrid modelling framework featuring an Anomaly Detection layer (Isolation Forest) for real-time edge monitoring and an RUL regression engine (CatBoost) to predict remaining brake pad life, achieving a 65% improvement in predictive accuracy for failure detection. Deployed scalable micro-services using Docker and Kubernetes to maintain high-availability inference pipelines, ensuring seamless integration with the HealthHub monitoring dashboard for fleet-wide health alerts.
Data Analyst
UDAAN
Jan 2020 - Apr 2021
Designed robust data pipelines using advanced SQL (CTEs, window functions, and complex joins) to process and centralize large-scale B2B e-commerce datasets, improving data accessibility for downstream analytics. Optimized database performance and integrity by refining stored procedures, triggers, and query structures, reducing data latency and ensuring 100% accuracy for mission-critical business reporting. Developed interactive Tableau dashboards for sales performance and seasonal trends, optimizing marketing strategies and inventory allocation by 65%.
Data Analyst
PFS web Global Services
Jul 2017 - Dec 2019
Engineered end-to-end reporting pipelines by integrating Python and SQL to automate data ingestion from SharePoint, achieving an 80% reduction in manual processing time. Architected robust automation frameworks using Python for complex data consolidation, replacing legacy manual workflows and ensuring consistent, on-time delivery of mission-critical executive reports.
Education
B.Tech / B.E. in Electrical - KKR and KSR Institute of Technology and Sciences
- 2017 · Afghanistan
Certifications
No certifications added yet
Interested in this developer?
Profile Score Breakdown
Profile Overview
Availability Details
Visa Status
Citizen
Relocation
Open to Relocation
Skills (28)
Click a skill to find developers with the same skill
