Summary
Overview
Work History
Education
Skills
Projects
Accomplishments
Timeline
Generic
Vijay  Kumar M

Vijay Kumar M

Chennai

Summary

Established professional well-versed with designing strategic initiatives. Excellent communication and negotiation skills. Dedicated to implementing process changes to meet needs of organization.

Overview

8
8
years of professional experience

Work History

Quality Lead, Rufus and Alexa Shopping

Vijay Kumar
Chennai
07.2024 - Current
  • Saved $202K per year by creating a project to improve efficiency in the ML space by identifying redundant annotations being done in the process at different sites and deploying a solution.
  • Reduced Turnaround Time (TAT) by 50% (from approximately 40 days to approximately 20 days) for the D-team goal process, reducing PL bandwidth consumption from 9 to 5 hours per week, and improving quality from 67% to 93% in nine weeks.
  • Managed a diverse team of over 15 members, overseeing more than five operational processes. Provided ongoing support for their monthly initiatives, mentored individuals to foster their professional development, and successfully facilitated the promotion of six team members to the next level.
  • Utilized data-driven insights and analysis to influence category leaders and stakeholders in strategic decision-making.
  • Managed operational escalations and processes, while spearheading the implementation of technology solutions, process improvements, and data-oriented analysis to achieve operational and business objectives.
  • Communicated with stakeholders across multiple lines of business regarding operational milestones, process changes, and escalations.
  • Demonstrated a proactive approach in identifying process gaps, and effectively collaborated with stakeholders to propose solutions, leveraging subject matter expertise, and providing guidance to peers and junior team members.

Senior Data Associate

Vijay Kumar
Chennai
12.2019 - 06.2024
  • Reduced the annotation disagreement between two teams from 8% to 2% in six weeks.
  • Created a checklist for onboarding a new process where our collaboration with the SE team lacked a structured approach to the workflow creation process, leading to dependency issues, and achieved a 9-hour weekly save.
  • Identified and drove process improvements, as well as surfaced opportunities for leveraging technology to streamline processes, enhance tools, increase productivity, and reduce defects.

Data Associate

Vijay Kumar
Chennai
09.2017 - 12.2019
  • Acted as a Subject Matter Expert for all processes in the first year of the team's creation, trained over 50 successful associates, and created several process and tool improvements.
  • Created a SmartCheck guide to avoid manual oversight errors when processes were done during annotations.

Education

Bachelor of Science - Computer Science

Rajalakshmi Engineering College
Chennai
05-2017

Skills

  • Program Management
  • Stakeholder Management
  • Strategic Thinking
  • Data Analysis
  • Advance Excel
  • Amazon QuickSight
  • Hubble Workbench

Projects

CX Defect Reduction & Catalog Quality Improvements

  • Led the Root Cause Analysis (RCA) of customer experience (CX) failures related to catalog inconsistencies in the DAWN process, successfully reducing the defect rate from 90% in Week 15 to 3% by Week 29
  • The initial RCA SOP provided by the product team was a basic 4-step draft, which I expanded and restructured into a comprehensive 16-step SOP
  • I partnered with Selection Monitoring (SM) and Global Catalog Operations (GCO) teams to fix catalog quality bugs across and title fixes yielding 2,582 bps (69% defect rate in WK15 improved to 44% in WK19.
  • · Helped Identified PR and legal vulnerabilities necessitate robust data validation processes and comprehensive model performance evaluation. These measures are critical for maintaining organizational compliance, safeguarding reputation, and mitigating potential liability exposure.
  • · SM/GCO/OPTIMA Catalog quality disagreement calibration: Partnered with SM/GCO team to calibrate on SOP ambiguities through bi-weekly defect calibration sessions . We’ve taken up an internal goal of achieving less than 5% disagreement rate for the defects reported by OPTIMA (Current Status: Trending GREEN for WK 18 at 2%).
  • · Catalog Inconsistency Defect Metric Revisions:Transitioned from ASIN-level to attribute-level defect reporting, enabling micro level (attribute) defect measurements with owners for each attribute. This changed the baseline from 60% defect rate at an ASIN level to 6.1% P0 defect rate at an attribute level. This approach has empowered the catalog team to set specific goals at an attribute level (2% defect rate goal for image and size, 5% defect rate goal for all other P0 attributes), with dedicated owners (CSS, SM teams) driving improvements for each attribute metric.
  • Starfish defect triaging:Starfish defect categories were expanded to provide comprehensive break up of root causes, accelerating the fixes at SF team. SF defects were expanded from 1 defect type (Starfish hallucination) to 4 defect types (Scraper defect, Incorrect starfish contribution, Redrive required, Reconciliation hint along with redrive required). Each defect type has a system owner (e.g.: CSS team owns incorrect starfish contribution) who reviews the defects and makes fixes within 10 business days

Onboarding project super fresh

  • · In Data Validation team received request from Science team to on-board super fresh process for measuring the model accuracy of fact check worthiness of claims. I spear headed the entire process on-boarding, where he worked with SDE team to understand the E2E requirement.
  • · I also validated the SOP created by the SOP specialist team and shared 10+ updates which has been revised in the final SOP. I later trained 40+ DAs in Data Validation/Quality team and helped in addressing 20+ ambiguous scenarios (Ex. vague claims, claims not fact check worthy etc.).
  • · Additionally, i submitted detailed requirement with SWOT team for workflow creation, data ingestion, and output processing. As the process was relatively new, the quality score was hovering around 87%. I was instrumental in conducting dive deep on the errors reported and identified 80% issues in annotating “Vague claim”.
  • · To mitigate, he conducted refresher sessions and assessments for the Ai-Trainers specifically focusing on addressing vague claim ambiguity which led to quality improvement from 87% to 99.2%. Due to all his effort, we were able to deliver ~9.3K volume as per the initial commitment with 99.2% quality.

Accomplishments

  • Feedback for stakeholders ""Great progress with reducing Catalog defects attributed to Starfish!” - Abraham Ray (Sr. Principal TPM - Dawn),Sanjay Malhotra (Principal TPM - Dawn),Amol Gupta(Principal TPM -CSS),Rahul Tamaskar (Sr. Manager - Product, Selection Monitoring)
  • Received a "Mount Mover" (R&R) for impactful contributions during Q4 2024.

Timeline

Quality Lead, Rufus and Alexa Shopping

Vijay Kumar
07.2024 - Current

Senior Data Associate

Vijay Kumar
12.2019 - 06.2024

Data Associate

Vijay Kumar
09.2017 - 12.2019

Bachelor of Science - Computer Science

Rajalakshmi Engineering College
Vijay Kumar M