AI Engineer · Motorsports · Brooklyn, NY

KAZI SADIA
ANOWAR

AI Engineer × Race Strategy Intelligence × Real-Time Telemetry

Computer Science and Artificial Intelligence student building agentic AI systems, real-time telemetry pipelines, and LLM-powered applications deployed to production. I work at the intersection of machine learning and motorsport, where milliseconds and model accuracy both decide the outcome.

VIEW PROJECTS → DOWNLOAD RESUME
CAREER TELEMETRY · LIVE FEED
0+
Telemetry frames processed
0+
Race laps ingested
0
Production AI systems
0
F-Student AI Judge · Silverstone
01 Projects 5 SHIPPED · CLICK TO EXPAND
RaceCast
Real-time Formula 1 telemetry analytics platform streaming live race data sub-second.
F1 · TELEMETRY
PythonFastAPIPostgreSQL RedisWebSocketsNext.js DockerFastF1
+ READ THE BUILD 84,000+ FRAMES
  • Architected and deployed a production-grade real-time telemetry platform ingesting 84,000+ telemetry frames and 1,400+ race laps from live Monaco 2025 race data.
  • The hard part was streaming: my first version pushed every frame to the client and choked, so I added Redis as a buffer and batched frames intelligently to keep it real-time without overwhelming the connection.
  • Designed REST APIs, WebSocket streaming pipelines, and cloud ingestion services with uptime monitoring and CI/CD deployment across Vercel, Render, and Neon.
↗ GitHub
PitWall AI
An LLM-powered race strategy agent that predicts pit windows and explains its reasoning.
F1 · AGENTIC AI
LangGraphAnthropic APIXGBoost Monte CarloFastF1PostgreSQLFastAPI
+ READ THE BUILD 1,000+ SIMS
  • Building an LLM-powered strategy agent that runs 1,000+ Monte Carlo race simulations and generates pit window recommendations with confidence scores and natural-language reasoning.
  • Developing an XGBoost tire degradation model trained on 3 seasons of FastF1 data, targeting accuracy above the published 70.59% F1 AI research benchmark.
  • Designed a four-node LangGraph pipeline: telemetry ingestion → Monte Carlo simulation → LLM reasoning synthesis → structured strategy output.
↗ GitHub
PitWall Sim
Backtestable Monte Carlo race strategy simulator validated against real race outcomes.
F1 · SIMULATION
PythonFastF1f1db Monte CarloNumPyPandas
+ READ THE BUILD VALIDATED
  • Built a backtestable Monte Carlo simulator modelling tire degradation, safety car probability, and undercut windows across historical F1 seasons.
  • Generated probabilistic strategy outputs with confidence intervals across 1,000+ simulated scenarios per session, validated against known race outcomes.
  • Designed a reproducible engine with configurable compound, circuit, and weather parameters for multi-strategy comparison.
↗ GitHub
Zenith
Multi-agent AI career intelligence platform with live job data and explainable recommendations.
AI · AGENTS
LangGraphOpenAI APIAnthropic API FastAPIReactAdzuna API
+ READ THE BUILD 100+ CATEGORIES
  • Engineered a multi-agent LangGraph system ingesting live job listings via Adzuna API, scoring role fit with LLM evaluation frameworks.
  • Generated explainable career roadmaps across 100+ job categories with a transparent AI scoring engine.
  • Implemented prompt engineering patterns for consistent agent output quality and structured evaluation pipelines.
↗ GitHub
Cleaf
Five-agent AI orchestration engine automating healthcare operational workflows.
AI · HEALTHCARE
LangGraphFastAPINext.js PostgreSQLRedisPython
+ READ THE BUILD 5 AGENTS
  • Engineered a five-agent LangGraph orchestration engine that autonomously detects missed visits and triggers escalation workflows.
  • Developed caregiver recommendation logic using real-time scheduling and availability data.
  • Built operational intelligence dashboards tracking service risk and resolution metrics; automated reporting pipelines.
↗ GitHub
02 The Stack WHAT I BUILD WITH

AI & Machine Learning

LangGraphLLMsRAGAI AgentsAgentic Workflows Prompt EngineeringAI EvaluationModel DeploymentFine-tuning MLOpsNLPVector DatabasesXGBoostPyTorchScikit-learn

Motorsports & Data

F1 Telemetry AnalysisFastF1Race Strategy Modelling Tire Degradation ModellingMonte Carlo SimulationTime-Series Analysis Real-Time PipelinesPerformance Analytics

Languages & Frameworks

PythonJavaSQLTypeScriptJavaScriptMATLAB FastAPIReactNext.jsNode.jsPandasNumPy

Infrastructure & Data

DockerAWSPostgreSQLRedisREST APIsWebSockets GitCI/CDPower BITableauCloud DeploymentSystem Design
03 Experience THE TRACK RECORD
JAN 2024 — PRESENT
Operations & Data Analyst
Marks Home Care · Brooklyn, NY

Built KPI dashboards monitoring 200+ caregiver schedules in SQL, Python, and Power BI, reducing manual reporting time by 30% through scripted data pipelines.

OCT 2023 — DEC 2023
Pharmacy Technician
Newkirk Pharmacy · Brooklyn, NY

Processed prescriptions and maintained data integrity across patient, prescription, and inventory systems in compliance with healthcare regulations.

FEB 2021 — JUL 2023
Research Assistant
Ujjibon · Bangladesh

Built data analysis pipelines across research projects; synthesized findings into strategic recommendations for evidence-based decisions.

◆ SELECTED · RECOGNITION

AI Statics Judge — Formula Student 2026

Selected to the international AI judging panel at Silverstone, UK — one of the most competitive applied-AI environments in motorsport.

◆ EDUCATION

B.S. Computer Science — AI

New York Institute of Technology · Expected Dec 2028.

LET'S BUILD SOMETHING FAST

Start your engines.

Open to AI engineering roles, motorsports data positions, and internships. If you're building at the edge of real-time AI, let's talk.

✉ sadiaanowar20@gmail.com ↗ GitHub ↗ LinkedIn