Computer Engineering @ University of Michigan

Tanish
Dalal.

|

Building ROS 2 robotics systems, telemetry platforms, and ML-driven infrastructure that ships into real test environments.

scroll

About

I'm a Computer Engineering student at the University of Michigan (graduating 2028), building at the intersection of robotics, machine learning, and full-stack software.

I'm currently a software engineering intern at Candela Technologies, working on ROS 2 autonomy tooling, wireless-network test automation, and multi-sensor robotics infrastructure for client deployments.

Previously, I built EV-charger telemetry and predictive-maintenance systems at ADNOC Distribution and co-founded Michigan's Robomasters team, where I lead computer vision and autonomy work on ROS 2 robots.

3.83

GPA

Dean's List · University Honors

2

Internships

Candela · ADNOC

3

Featured Projects

Resume-aligned work

Experience

SWE

Software Engineering Intern

Candela Technologies

Jul 2026 – Present

Ferndale, WA (Remote)

  • Integrated 2D LiDAR and IMU data with ROS 2 Cartographer through Ansible-driven deployments for autonomous surveying across 6 client environments
  • Built multithreaded Python automation pipelines against LANforge REST APIs across 15 access points, cutting test turnaround from 3 days to 4 hours
  • Synchronized robot telemetry with RSSI and latency metrics using ptp4l/phc2sys hardware time sync to map network performance to spatial coordinates
  • Resolved ROS 2 multi-subnet connectivity bottlenecks by tuning CycloneDDS discovery settings and analyzing Wireshark captures, reducing downtime 30%
PythonROS 2CartographerAnsibleCycloneDDSWireshark
Backend

Software Engineering Intern

ADNOC Distribution

EVCI & H2 Department

May 2025 – Jul 2025

Abu Dhabi, UAE

  • Built an OCPP 1.6 telemetry ingestion service in Python and FastAPI for 400+ EV chargers with 7k+ points per minute at under 100 ms p99 latency
  • Engineered an XGBoost predictive-maintenance model from 40+ rolling telemetry features, reaching 0.78 PR-AUC and detecting failures 30 minutes before shutdown
  • Wrote pytest regression coverage for malformed payloads, charger reconnections, duplicate events, and database failures, reducing manual validation time 60%
PythonFastAPIWebSocketsTimescaleDBXGBoostpytest

Projects

Aug 2025 – PresentRobotics

Michigan Robomasters

Computer Vision Lead & Co-Founder

Lead a 6-member team building a ROS 2 autonomy stack for Robomasters competition robots across computer vision, tracking, and navigation.

  • Trained YOLO26n on 8k+ labeled armor-plate frames to 0.58 mAP@50:95 and integrated solvePnP pose estimation at 45 FPS on Jetson Orin Nano
  • Implemented a constant-velocity Kalman filter for lead compensation, reducing angular tracking error 28% in autonomous shooting tests
  • Built sentry-robot autonomy with FAST-LIO2, Livox MID-360 LiDAR, and Nav2 in ROS 2
  • Achieved a 90% collision-free route-completion rate in Gazebo and on the physical robot
C++PythonROS 2OpenCVYOLONav2FAST-LIO2Jetson Orin
View on GitHub
Aug 2024 – PresentRobotics

Michigan Mars Rover

Robotic Arm Controls Developer

Built closed-loop inverse-kinematics and workspace analysis tooling for a 5-DOF rover arm in a 30 Hz ROS 2 control loop.

  • Engineered a closed-loop Cartesian IK controller that improved target alignment accuracy 20%
  • Integrated geometric IK solvers into a 30 Hz ROS 2 velocity pipeline with joint-limit checks
  • Achieved under 1 cm convergence while reducing motor saturation 30%
  • Built a Python reachability visualizer with dexterity scoring and draggable RViz obstacles, cutting validation time 50%
C++PythonROS 2RVizEigenNumPy
View on GitHub
Sep 2025 – PresentFull-Stack

Algorithm Practice Platform

Sole Developer

Real-time multiplayer algorithm battle platform with secure code execution, live matches, and automated interview-style problem generation.

  • FastAPI and WebSockets backend supporting 50+ concurrent matches with <150 ms median latency
  • Secure code execution engine supporting 3 languages and a C++ assembly viewer
  • Big-O analysis via log-log regression and TF-IDF + k-means approach classification
  • ELO-seeded matchmaking, 8-player tournament brackets, and live rating updates via SQLite
  • AI problem and interview generator with 96% validation success and <1.2 s responses
ReactFastAPIPythonJavaScriptSQLiteWebSockets
View on GitHub

Skills

Languages

C/C++PythonJavaScriptSQLJavaAssemblyVerilogMATLAB

ML & Vision

PyTorchUltralytics YOLOOpenCVXGBoostscikit-learnpandasNumPy

Robotics & Data

ROS 2Nav2CartographerFAST-LIO2RVizGazeboEigenCycloneDDS

Frameworks & Systems

FastAPIReactWebSocketsREST APIsPostgreSQLTimescaleDBSQLiteAnsibleWiresharkpytestGitHub ActionsLinuxSTM32Jetson Orin

Contact

Let's build something.

Whether it's a robotics project, internship opportunity, or something I haven't thought of — my inbox is open.

Tanish Dalal · University of Michigan · Class of 2028