HE.HAO
Stack

Technical Map & Working Profile

A calibrated view of how I work across optimization, autonomous systems, applied ML, and platform engineering, with emphasis on technical depth, executable autonomy, and the system boundaries that turn methods into working platforms.

01 · Proficiency
Proficiency:
Expert
Advanced
Intermediate
Learning
02 · Skill Groups
Distributed Optimization
Gradient tracking, ADMM, Push-Sum, decomposition, and async/distributed variants
Nonlinear Optimization
Convergence analysis, nonconvex structure, KKT conditions, and saddle-point behavior
Multi-Agent Systems
Consensus, coordination, swarm guidance, coverage, and formation-aware design
Continuous-Time & Hybrid Optimization
Settling-time design, equivalent circuits, dynamical-system interpretation, and hybrid trajectory generation
Nonconvex Optimization
Constrained nonconvex formulations, reformulation strategy, saddle structure, and decomposition-aware design
Guidance, Tracking & Control-Oriented Modeling
Executable motion design, tracking logic, and control-oriented formulation
Planning & Guidance
Path following, guidance boundaries, mission-level reasoning, and route-level autonomy
Path Planning
Potential fields, lattice-style reasoning, and planning-stage autonomy design
Trajectory Generation
Smooth reference generation for executable vehicle motion
Simulation & Experimentation
MATLAB-based control simulation, surveillance studies, and autonomy experimentation
Navigation Stack Architecture
Mission -> planning -> estimation -> guidance -> control boundaries and system interfaces
State Estimation & Sensor Fusion
EKF, UKF, GNSS/IMU/DVL integration, and navigation fusion
Supervised Learning
Classification, regression, evaluation, and error analysis
Computer Vision Systems
Detection pipelines, inference services, and operator-facing CV workflows
PyTorch
Custom training loops, inference integration, and model-facing engineering
Signal Processing
Denoising, smoothing, and calibration-first pipelines for noisy observations
Evaluation & Ensembling
Cross-validation, model comparison, ensemble design, and competition iteration
Gradient-Boosted Trees
XGBoost, LightGBM, and structured-data modeling
API & Service Architecture
Shared cores, contract boundaries, and modular backend organization
FastAPI
Service assembly, router design, tool boundaries, and inference APIs
React
Operator-facing interfaces, dashboards, and interactive workflow surfaces
TypeScript
Typed frontend architecture and contract-driven development
PostgreSQL
Catalog-backed services, schema design, and operational data models
Docker
Reproducible development, local deployment, and service packaging
Python
Optimization, ML, backend services, and data workflows
MATLAB
Control design, simulation, and rapid prototyping for research
CVX / Optimization Tooling
Convex modeling, solver-based prototyping, and experiment design
C++
Systems-level implementation and performance-sensitive modules
Git
Versioned research and production workflows
LaTeX
Papers, thesis writing, and technical documentation
03 · Preferred Stack
Preferred Stack

For most current work, the research side leans on Python + MATLAB + solver-driven optimization tooling for optimization, simulation, and technical experimentation. The autonomy side emphasizes estimation, planning, guidance, executable motion design, and navigation-stack boundaries, while the platform side is shaped by FastAPI + React + PostgreSQL + Docker for operator-facing tools, CV runtimes, and service-backed workflows. I choose tools based on system needs, execution constraints, and clarity of architecture rather than breadth for its own sake.