Resume

Shlok Thakkar

shlokthakkar1806@gmail.com · linkedin.com/in/shlok-thakkar · github.com/shlok1806

Education

2023 – 2027

B.S. Computer Science + Economics

University of Illinois Urbana-Champaign · Champaign, IL

Minor in Statistics · GPA: 3.97/4.00 · Expected May 2027

Distributed Systems · Algorithms · Machine Learning · Database Systems · Operating Systems · Econometrics

Technical Skills

LanguagesGo · C/C++ · Python · Swift · TypeScript · SQL · Java · R
FrameworksFastAPI · Flask · SwiftUI · Combine · SvelteKit · LangChain
DatabasesPostgreSQL · Neo4j · Supabase
Cloud & DevOpsAWS (Bedrock, S3, EC2) · Azure Blob · Docker · Git · CMake
LibrariesNumPy · Pandas · SciPy · Scikit-learn · lxml · OpenAI API

Experience

Apr 2026 – Present

Undergraduate Research Assistant

Parallel Programming Laboratory, UIUC · Champaign, IL

  • —Extended a NumPy-based distributed array abstraction on Charm++ by implementing custom operators, relational/logical operations, and broadcasting for high-performance parallel computations.
  • —Contributing to Reconverse, a ground-up rewrite of Charm++'s core runtime, by redesigning core APIs to improve developer ergonomics and long-term extensibility.
May – Dec 2025

Undergraduate Research Assistant

Department of Finance, UIUC · Champaign, IL

  • —Developed a custom NLP pipeline combining TF-IDF vectorization, Levenshtein distance, and fuzzy matching to link firms across 10M+ record TED and ORBIS databases for economic research.
  • —Preprocessed 10M+ records through name normalization, deduplication, and text cleaning to improve matching signal quality before algorithm application.
  • —Achieved a 63% firm-level match rate across TED and ORBIS, enabling large-scale empirical analysis for Prof. Tatyana Deryugina's research.
Aug 2024 – Present

Software Developer

Disruption Lab @ UIUC · Champaign, IL

Fortune 500 financial services client · Jan – Apr 2026

  • —Built a Python-based XSD parser for a 44-jurisdiction tax schema pipeline using recursive lxml traversal and XLSX business rules ingestion.
  • —Wrote 43 unit tests covering XSD fixtures and business rules mappings, ensuring data integrity before downstream pipeline integration.
  • —Expanded 2023–2024 schema diff coverage from 7 to 27 states (+286%) by implementing version-based XSD deduplication.

DSRS · Aug – Dec 2024

  • —Developed a REST API using Python (Flask) and Neo4j, enabling 40+ users to visualize project-resource dependency networks in real time.
  • —Improved platform interactivity by implementing a node update endpoint and optimizing Cypher queries, converting a read-only graph into a fully editable tool.
May – Aug 2024

Software Development Intern

IQM Corporation · Ahmedabad, India

  • —Built a FastAPI service with JWT authentication and rate limiting to interface with AWS Bedrock (Titan) via LangChain, delivering political personality analysis to 3 internal teams.
  • —Engineered a LangChain prompt pipeline identifying the 10 closest personality matches per political figure, persisting structured inference outputs to AWS S3 as CSVs.

Projects

Apr 2026
raft-kvGo
  • —Implemented Raft consensus from scratch in Go — leader election, log replication, snapshotting, and fast log backup — supporting a fault-tolerant distributed key-value store.
  • —Engineered exactly-once semantics via client-side request deduplication and per-index pending channels, ensuring no committed write is lost or applied twice.
  • —Validated correctness under failure by writing chaos tests that kill and reconnect random nodes over 30 seconds, verifying zero data loss across concurrent clients.
Mar 2026
Limit Order BookC++20, CMake, Catch2
  • —Engineered a thread-safe C++20 limit order book with 5 order types (GTC, FOK, FAK, GFD, Market) and FIFO price-time priority matching.
  • —Achieved O(1) cancellation by storing per-order iterators into price-level linked lists, eliminating linear scans across resting orders.
  • —Built an incremental LevelData cache reducing FOK pre-check complexity from O(N) to O(P), a ~500x improvement for sparse price-level books.
Feb 2026
Scroll RoyaleSwiftUI, Combine, AVKit, Supabase, PostgreSQL
  • —Built a 1v1 competitive iOS doomscrolling game at HackIllinois using SwiftUI and Supabase with real-time telemetry-driven scoring.
  • —Implemented matchmaking and live leaderboard sync over Supabase PostgreSQL RPC functions with JWT auth and proactive token refresh.
  • —Designed a TTL-based feed cache with stale-while-revalidate and adaptive score polling to minimize cold-start latency.
2025
Blueprint QAFastAPI, SvelteKit, Claude API, PostgreSQL, Docker, Azure
  • —Built an AI-powered construction drawing QA tool using FastAPI and Claude's multimodal API, automatically flagging missing tags, dimension mismatches, and unlabeled elements across uploaded PDFs.
  • —Engineered a multi-stage OCR pipeline using pdf2image and pytesseract to extract text from engineering drawings, feeding structured page data into Claude for per-page issue detection.
  • —Designed a pluggable storage backend supporting both local filesystem and Azure Blob Storage, and containerized the full stack in a multi-stage Docker build serving the SvelteKit SPA and API from a single origin.
2025
FeeLensNext.js, TypeScript, Stripe, Supabase, Claude API
  • —Built a Stripe fee analytics dashboard in Next.js that surfaces hidden costs — Amex premiums, international surcharges, refund fee retention, and dispute losses — giving startups a true effective rate breakdown by card brand, geography, and payment method.
  • —Engineered a Stripe data pipeline using paginated balance_transactions fetching with concurrent charge enrichment via p-limit, computing per-transaction fee leakage and persisting structured records to Supabase with row-level security.
  • —Integrated Claude API to generate ranked, data-driven fee optimization recommendations cached by SHA-256 input hash, with AES-256-GCM encryption of Stripe keys at rest.
2025
VibeSafeTypeScript, GitHub Actions, Claude API
  • —Built a GitHub Action in TypeScript that uses Claude to automatically scan pull requests for exposed secrets, insecure configs, and vulnerable dependencies, posting structured security findings as PR comments.
  • —Implemented multi-category static analysis covering hardcoded API keys, high-entropy strings, permissive CORS configs, and missing HTTP security headers across any language codebase.
  • —Designed the action as a reusable workflow with configurable severity thresholds and ignore patterns, enabling teams to enforce security gates on every PR without modifying existing CI pipelines.