A curated list of freelance projects completed since 1st year of college, highlighting cross-domain problem solving, dashboard building, data wrangling, and automation.
Filter projects
Select one or many to narrow the list
Country
Pick one or more countries
Tech Stack
Pick one or more tools
Architected a low-latency RAG pipeline for internal document intelligence. Solved the 'hallucination at scale' problem by implementing hybrid search and custom reranking schemas. Turned a slow search tool into a high-performance production engine.
Business Impact
Reduced document retrieval latency by 85% (45s to 3s). Scaled to 1,200+ MAU with 92% search accuracy, saving $12k/month in manual overhead.
Eliminated manual support bottlenecks by building a deep CRM-integrated AI agent. Focused on reducing state-machine inefficiencies and ensuring 99.9% reliability in order processing calls.
Business Impact
Automated 68% of support volume. Achieved 88% user retention. Contributed to a leaner support team and faster response times.
Optimized NLP pipelines for a high-traffic support system. Successfully migrated messy legacy graph logic to a clean, production-ready RASA architecture designed for 10x scale.
Business Impact
45% reduction in compute costs. Dropped call center logs by 12% via improved intent classification (89% -> 96%). Saved 60+ engineer hours monthly.
Built a high-throughput sentiment analysis system for Shopify/Etsy stores. Focused on turning raw review noise into actionable product roadmaps for high-growth e-commerce brands.
Business Impact
Enabled real-time sentiment alerts, reducing response time by 70%. Enabling proactive issue resolution and boosting customer satisfaction metrics.
Re-engineered a bloated legacy BI system into a lean ETL pipeline. Solved the problem of fragmented data silos by creating a single source of truth for CRM metrics.
Business Impact
Automated 20+ hours of manual weekly reporting. Identified ยฃ1000 in churn risks through new predictive retention dashboards.
Architected a multi-outlet NLP extraction layer to validate complex market correlation hypotheses. Engineered resilient scrapers to capture sentiment and narratives across 15+ global news outlets, enabling data-driven factor analysis for a London finance study.
Business Impact
Successfully validated multiple market hypotheses correlating news sentiment with price action. Streamlined the factor analysis loop, reducing a 300+ hour manual research effort into a week of automated data crunching.
Built a scalable analytics layer for regional sales data. Focused on reducing the time-to-insight for regional managers through automated data cleaning and visualization.
Business Impact
Reduction in regional reporting lag by 60%. Enabled data-driven territory expansion that resulted in 12% MoM revenue growth.
Engineered high-volume Logstash pipelines to ingest ticket transaction data. Optimized data flow from raw logs to executive-ready BI dashboards.
Business Impact
99.9% data integrity across 1M+ monthly transactions. Reduced business intelligence turnaround from days to minutes.
Designed and deployed health monitoring ETL pipelines. Focused on data precision and reliability for critical health analyst workflows.
Business Impact
Improved monthly reporting efficiency by 40%. Ensured 100% compliance in health trend tracking for UK-based health analysts.
Built a resilient, periodic web scraping engine for Facebook Marketplace and Zillow. Engineered custom proxy rotation and captcha solving to ensure 24/7 data availability.
Business Impact
Captured daily pricing trends for 1000+ listings. Provided the data edge needed for a $500k real estate investment fund's market entry.
Developed a Kibana-based debugging environment for massive system test logs. Focused on reducing search friction for internal QA teams to speed up software release cycles.
Business Impact
Reduced test-cycle troubleshooting time by 40%. Automated root-cause analysis for 25% of recurring system failures.
11+ Infrastructure Deployments
From Day 0 Prototype to Planning to MVP to real users.