Lorenzo Kamanzi

Computer Science × Finance

Co-Founder & Software Engineer at Creo. Rutgers University, Class of 2028.

Get in TouchView Resume
Finance & quant resume
About Me

Building at the Intersection of Software & Finance

I'm a double major in Computer Science and Finance at Rutgers University–New Brunswick, graduating May 2028. With one co-founder I built Creo, a startup validation platform with 984 signups: a weighted scoring engine and a competitor check that screens ideas before an LLM analysis.

In summer 2025 I led an Odoo ERP rollout at DIKAM Fashion, a garment manufacturer in Kigali: I wrote a custom manufacturing module, migrated ten years of Excel books into SQL, and helped cut month-end close from two weeks to four days. That same summer I analyzed 2.4M+ NYC crash records for Columbia's Northeast Big Data Innovation Hub.

On my own time I build things like TradeDesk, a stateful LLM chat service on Cloudflare Workers and Durable Objects; market-data, a Java and PostgreSQL service for prices and indicators; and a quantitative trading simulator with a backtesting engine and risk layer. I'm looking for software engineering internships and roles where engineering meets finance.

Quick Facts

984

Signups on Creo, the startup I co-founded

Top 5%

HackRU Finalist, Fall 2025 (142 projects)

3.8

GPA, Dean's List

Scholar

All-Star Code

Education

Rutgers University – New Brunswick

B.S. in Computer Science and Finance (double major) · GPA 3.8/4.0

Expected May 2028

New Brunswick, NJ

  • Honors: Dean's List, HackRU Finalist (Fall 2025, top 5% of 142 projects), All-Star Code Scholar
  • Activities: NSBE, CodePath, USACS
  • Coursework: Data Structures & Algorithms, Computer Architecture, Discrete Math, Probability
Selected Projects

Building Systems,
Creating Value

Edge Backend · Apr 2026

TradeDesk – Stateful LLM Chat Service on the Edge

Ticker-aware trading research chat on Cloudflare's edge stack, with one Durable Object per conversation.

  • One Durable Object per session: single-threaded, strongly consistent conversation state with no locking in app code
  • Answers grounded through Workers AI tool calls to the market-data API; a post-check flags any price not in the data
  • Streams Llama 3.3 70B over SSE; a tee'd copy is persisted to DO storage and D1, indexed on (ticker, created_at)
  • Per-IP rate limits, a daily D1 budget and token caps; 548 vitest tests across the Workers pool and jsdom
  • Vite front end: DOMPurify-sanitized markdown, voice input, history sidebar, Lightweight Charts with the cited levels
TypeScriptCloudflare WorkersDurable ObjectsD1 (SQLite)Workers AIVite
View on GitHub
Quant Finance · Dec 2025 – Present

Quantitative Trading Simulator

Backtesting platform for systematic trading strategies, with a risk layer and synthetic market data.

  • Backtesting engine: pluggable strategies (MA crossover, RSI, breakout with ATR stop), Sharpe, CAGR and drawdown metrics
  • Risk layer (position caps, stop-losses, drawdown halt) vs. an unmanaged baseline; Flask API and React/TypeScript UI
  • Profiled the bar loop and precomputed indicators: 25.8 s to 0.12 s on 32,625 rows, results byte-identical in CI
  • PostgreSQL schema under Alembic (NUMERIC money, TIMESTAMPTZ, CHECKs); drawdown and Sharpe re-checked in SQL
  • Seeded Markov regime-switching GBM generator; 430 pytest and 32 frontend tests in CI on Python 3.11 and 3.12
PythonFlaskPostgreSQLReactTypeScriptDockerpytest
View on GitHub
Backend / SQL · Oct 2026

market-data – Market Data Service in Java and SQL

Ingests daily US equity bars and splits into PostgreSQL and serves prices, indicators and levels over a REST API.

  • Java 25 and Spring Boot 4 with hand-written SQL through JdbcClient (no ORM); Flyway migrations on PostgreSQL 18
  • Idempotent ingestion (ON CONFLICT upserts, retries with backoff); split-adjusted prices in a window-function view
  • SMA, volatility, ATR, 52-week range and pivots in SQL; EXPLAIN-driven view rewrite: last 100 bars 7.62 ms to 0.59 ms
  • REST API with RFC 9457 errors, keyset pagination, ETags, per-IP rate limits and scoped API keys; React chart explorer
  • JUnit, Testcontainers and WireMock tests; k6 load test at 200 req/s: p95 19.5 ms (local run, synthetic data)
JavaSpring BootPostgreSQLFlywayTestcontainersReactDocker
View on GitHub
Data Engineering · Oct 2026

ERP Migration Demo – Excel Books to PostgreSQL

A reconstruction, on synthetic data, of the approach behind the DIKAM migration: messy Excel books into a normalized schema.

  • Seeded generator writes 42 messy workbooks (mixed date formats, RWF and USD amounts, misspelled suppliers) plus ground truth
  • Python and SQL ETL into PostgreSQL 18: constraints, a rejects table with reason codes, window-function de-duplication
  • Supplier matching scored against the ground truth: precision 1.000, recall 0.758, no wrong merges
  • Data-quality report: 21,120 rows in, 18,901 loaded, 530 duplicates merged, 325 rejected with a reason
PythonPostgreSQLSQLpandaspytestTestcontainers
View on GitHub
Experience

Professional Journey

Co-Founder & Software Engineer

Creo

Jun 2025 – Present

New Brunswick, NJ

Startup validation platform with 984 signups and 25–50 daily active users.

  • Co-built Creo with one co-founder: Next.js/TypeScript, Supabase auth, GPT-4o returning JSON
  • Scoring engine: 28 weighted questions in six dimensions roll up to a 0–100 score with red/yellow flags and a stage label
  • Competitor check: a keyword pre-filter flags ideas in crowded markets and asks the founder for a differentiation statement
  • A rule-based validator screens vague differentiation claims before the GPT-4o competitor analysis runs
  • Reworked the scoring flow on beta feedback; onboarding drop-off fell 28% from the 45-user beta to the next cohort
Next.jsTypeScriptSupabaseGPT-4o

ERP Implementation Lead

DIKAM Fashion Ltd

May 2025 – Aug 2025

Kigali, Rwanda

Garment manufacturer; founder in Africa's Business Heroes 2025 Top 10.

  • Wrote an Odoo module in Python ("Manufacturing v2") that replaced the stock Purchase-to-Manufacturing flow
  • Migrated 10 years of Excel purchase, sales, payroll and inventory books into normalized SQL tables with Python
  • Reconciled inconsistent supplier names, date formats and currencies across the books before load
  • Led the Odoo rollout (production, inventory, procurement, payroll, finance); month-end close: 2 weeks to 4 days
PythonOdooSQL

Data Analyst Researcher

Columbia University – Northeast Big Data Innovation Hub

May 2025 – Jul 2025

New York, NY (Remote)

Borough-level study of NYC motor-vehicle crash density.

  • Analyzed 2.4M+ NYC motor-vehicle crash records in Python (Pandas) for a borough-level crash-density study
  • Built the cleaning and aggregation pipeline and the dashboards behind the team's published report
PythonPandasData Viz
Leadership

Treasurer

TWESE African Association

Sep 2024 – Present

Rutgers University

  • Manage the $40,000 annual budget of a 100+ member pan-African student organization
  • Primary liaison to the Rutgers University Student Assembly (RUSA) for supplemental funding and financial reporting

Outreach Chair

ColorStack

2025 – Present

Rutgers University Chapter

  • Organized a 75+ attendee networking event connecting members with engineers and internship openings
  • Recruit new members and bring recruiters and engineers to chapter events
Skills & Expertise

Technical Arsenal

Languages

JavaPythonC++TypeScriptJavaScriptKotlinDartPHPSQLHTML/CSS

Frameworks & Tools

Spring BootReactNext.jsFlutterFlaskPandasNumPypytestGitGitHub ActionsDocker

Cloud & Databases

Cloudflare WorkersSupabaseFirebase Firestore (NoSQL)PostgreSQLSQLite

AI-Assisted Development

Claude CodeCursorKiroCodex
10
Languages
11
Frameworks & Tools
5
Cloud & Databases
4
AI Dev Tools
Contact

Get in Touch

I'm looking for Summer 2027 internships in software engineering and quant or fintech engineering. Email is the fastest way to reach me.

Email MeLinkedInGitHub