Associate Consultant
Ernst & Young India
05.2022 - Current
- Designed and deployed a success-profile generation algorithm, leveraging NLP-based parsing of job descriptions and a normalized skills ontology, implemented custom data cleaning and deduplication pipelines that reduced skill redundancy by 35%, and achieved 85% skill mapping accuracy, decreasing manual QA effort by 60%.
- Developed a time series analysis-based skill trend tracking algorithm for over 20,000 skills, applying statistical methods such as interpolation, anomaly detection, hypothesis testing, and deployed via Apache Airflow to produce quarterly labor market intelligence.
- Built robust, multi-source data ingestion pipelines using AWS, MongoDB, and SQL databases, implementing schema normalization, deduplication, and validation to standardize and structure data for 20+ downstream analytics pipelines; achieved 99.9% data reliability, improving data accessibility, and operational efficiency.
- Implemented asynchronous API interactions to execute high-volume, concurrent endpoint requests, optimizing network utilization, reducing API response time by 40%, and enabling real-time data processing pipelines that improved system throughput by 65%.
- Designed and deployed a personalized content recommendation engine for 100,000+ courses from 40+ providers, utilizing Apache Airflow to orchestrate workflows and drive a measured 35% increase in user engagement with learning pathways.
- Directed the analysis of over a billion data points to generate key business insights for comprehensive skill gap analysis across 20+ industries and 20,000+ unique skills, directly informing strategic investment in training programs.
- Developed interactive Q&A chatbot solutions, leveraging LangChain and large language models (LLMs), with a vector database for robust information retrieval from internal documents, improving internal knowledge access speed by 3x.
- Developing a novel Text-to-AI-Video algorithm (currently in prototype) designed to reduce content creation costs by an estimated 30% and scale video production capacity for internal training initiatives.
- Revamped content mapping logic by developing an optimized algorithm that reduced the processing time for catalog mapping from 6 hours to 5 minutes (a 98.6% efficiency improvement).
- Synthesized Employee Knowledge Reports for stakeholders by integrating and analyzing self-ratings, manager ratings, and test scores, improving training efficiency by 25%, and guiding prioritization for 100% of the annual training budget.
- Conducted comprehensive statistical analysis on complex labor market datasets, producing actionable insights that guided senior leadership's strategic decision-making and optimized resource allocation for talent development initiatives.
- Developed and maintained daily Python automation scripts to enforce rigorous data quality checks, achieving 100% data integrity and fully eliminating manual errors; collaborated with cross-functional teams to optimize workflows, saving over 20 hours of repetitive tasks per week.
- Collaborated with cross-functional teams to optimize system workflows, improve operational efficiency, and reduce process bottlenecks, resulting in faster project delivery and enhanced system utilization.