I'm Ankit — a Computer Science undergraduate from Guwahati, India, working across backend engineering, real-time platforms, and network security. I like problems where correctness and latency both matter.
Curious about how things break, before they do.
I'm a B.Tech Computer Science student at Assam Don Bosco University, currently in my final year. Most of my work sits at the intersection of backend systems and cybersecurity — building the services that move data, and then checking whether that data can be trusted.
That's shown up as kernel-level network monitoring with eBPF and XDP, real-time platforms handling hundreds of concurrent users, and data pipelines that clean and validate enterprise-scale records for cyber risk assessments. I'm equally comfortable in Python and Java, and I default to measuring things — latency, throughput, query time — rather than guessing.
Outside of coursework, I'm currently a Cybersecurity Track Analyst at PwC's Launchpad Program, and previously interned at C-DAC CINE, IIT Guwahati, building network security tooling.
Where I've been putting this to work.
Cybersecurity Track Analyst
- Applied Python and data libraries (Pandas, NumPy) to clean, transform, and validate thousands of structured records, supporting cyber risk and data integrity assessments for enterprise clients.
- Tested, debugged, and optimized solutions in Jupyter Notebook and VS Code, collaborating cross-functionally on enterprise-scale cybersecurity and data-system simulations.
Cybersecurity Intern
- Designed and implemented an eBPF- and XDP-based system to monitor desktop applications by analyzing TLS-encrypted network traffic, with low-overhead, kernel-level packet processing.
- Built a real-time security monitoring pipeline integrating Suricata, OpenTelemetry, Prometheus, and Grafana to collect, process, and visualize network telemetry and security events.
Two systems I've built end to end.
Both started from the same question: how does this hold up under real, concurrent, real-time load?
Quizzera
- Architected a real-time multiplayer trivia platform using Llama-3 for on-demand question generation, supporting 100+ concurrent sessions.
- Developed a Python microservice with regression-based difficulty adjustment to personalize gameplay based on user performance.
- Built a low-latency Socket.IO backend enabling sub-200ms real-time updates, concurrent game rooms, and persistent analytics.
Meteorological Data Analysis & Forecasting Platform
- Engineered a weather data platform processing 2TB of data daily, reducing query time by 30% through optimized indexing and partitioning.
- Built a time-series forecasting pipeline using autoregressive models to generate accurate 96-hour temperature predictions.
- Developed a high-performance REST API with interactive visualizations, enabling sub-second analytics and CSV data export.