|
Third-year CSE at IIT Jodhpur, drawn to the unglamorous half of a system. I build the layer other code depends on being correct: a rollback engine that makes writes to a live PostgreSQL database reversible, a system design tester that runs your architecture under load and injected failures before you build it, and a C++20 quant finance library whose hot loop has no branches in it at all. That usually means writing the interesting part myself rather than importing it — my own scheduler, my own inverse-query generation, my own AVX2 kernels. Slower than reaching for a framework, and the only way I've found to learn where a design actually breaks. Currently coordinating the Quant Club and building with DevlUp Labs. |
Looking for a Summer 2027 internship — software engineering or quantitative development. Reach me at pcshah2006@gmail.com |
ONITRO · C++20 AVX2 nanobind
Quant finance library closing the gap between QuantLib and ad-hoc Python. Branch-free, allocation-free kernels over struct-of-arrays; four doubles per instruction. Converged SIMD lanes are masked off instead of dropping to scalar. Architected it, lead-mentoring the team.
SimRUN · C++ DSL Discrete-event
Draw a system architecture, run it under load with failures injected, watch how it degrades before building any of it. 20-opcode component model over a timestamp-ordered scheduler. Built the simulation core and the topology DSL.
WEAVE-DB · Django FastAPI Kafka
Version control for a live PostgreSQL database. Inverse-query generation makes writes reversible; anchoring on the nearest snapshot either side of the target halves worst-case replay. Backend lead, team of four.
Custom Language & Interpreter · C Flex Bison
A language with its own syntax and datatypes. I wrote the lexer and the parser that builds the AST; the team wrote the tree-walking interpreter.
Maze Pathfinding Visualiser · C++17 SFML
Five search algorithms raced across generated mazes, with a benchmark screen comparing runtime, nodes explored and path length.
The Hive · Node pgvector Kotlin
Browser extension that classifies your current tab by vector search into a productivity score, with idle and minimum-time checks so the score can't be farmed.