Achieved $3M returns with 92% accuracy through deep learning and LLM-powered
trading systems with real-time ML pipeline
architecture.
Developed deep learning models using PyTorch and transformers for sentiment
analysis, deploying RAG architecture with
vector stores for real-time market analysis across 5K securities daily, fine-tuning
models with synthetic data
generation to augment training datasets.
Implemented A/B testing frameworks for systematic strategy optimization across
multi-model configurations, applied
causal inference to evaluate treatment effects and isolate true strategy performance
from market noise, establishing
comprehensive monitoring with continuous learning.
08/2024 - 08/2025Florida, USA
Achieved 10x opertaional efficiency improvement through end-to-end deep
learning powered incident management systems
development.
Developed statistical models for time series forecasting for incident prediction,
integrated with real-time alerting
system and executive dashboards, supporting real-time troubleshooting and
data-driven decision-making by the operational
team and stakeholders.
Built data engineering pipelines for log files with temporal patterns, processing
100GB daily logs via Splunk endpoints
for stabale and reliable data quality.
10/2022 - 05/2024California, USA
Improved data-driven decision-making efficiency by 3x through comprehensive
data analysis for Meta's global data center
capacity planning.
Built interactive and real-time Tableau dashboards visualizing hardware utilization
trends and capacity metrics across
global regions, enabling executives and operations teams to identify optimization
opportunities and make informed
resource allocation decisions.
Architected ETL and CDC pipelines processing billions of data points daily feeding
into Tableau dashboards, optimized
query performance by 180x enabling real-time capacity forecasts visualization for
executives to make informed resource
allocation decisions.
