Akash Dubey

Akash Dubey

I've been working in software for over 10 years 😉

Akash Dubey

Akash Dubey

Berkeley Heights, NJ | 212-814-3681 | akashdub06@gmail.com Portfolio | GitHub | LinkedIn

Summary

Engineer and Researcher at Rutgers University (CS + Math) with 10+ years of software experience. Specialized in AI/ML, Agentic Systems, Optimization, and High-Performance Computing. Architected systems serving 750+ users and led research for state/federal agencies.

Education

Rutgers University Honors College | New Brunswick, NJ Bachelor of Science in Computer Science and Mathematics | GPA: 3.97/4.0 | Expected May 2027

  • Relevant Coursework: Tensor Networks, Systems Programming, Algorithms, Data Structures, Statistics, Honors Calculus III & IV, Math Theory of Probability, Computer Architecture.
  • Honors: Phi Beta Kappa, Rutgers Honors College Scholar.

Experience

Rutgers Economics Labs (REL) | President / Research Director

Oct 2024 – Present

  • Leading development of the Rutgers Agentic Intelligence Lab, a general-purpose framework for structured economic analysis and reusable data pipelines.
  • Managing quantitative research for partners including NJDOL, NJDEP, NJBPU, Federal HUD, and U.S. Census Bureau.
  • Directed a team of 5 researchers analyzing the Urban Enterprise Zone program using econometric methods in Python/R.
  • Scraped and synthesized 20+ years of NJ state budget data to project economic funding variables.

Scarlet Sync | Founder / Lead Architect

Jan 2025 – Present

  • Built a modern Rutgers scheduling app serving 750+ users, supporting 2,500+ classes and 500+ degree options.
  • NSF I-Corps Participant: Completed the NSF I-Corps program for customer discovery and market validation.
  • Architected a Python/LLM data ingestion pipeline to reverse-engineer legacy university data.
  • Negotiating enterprise integration with Rutgers IT to replace legacy scheduling systems.

Lykke | Software Engineer

Sep 2025 – Present

  • Designing a multi-model AI pipeline for high-velocity data ingestion (text, voice, docs) for agentic RAG applications.
  • Engineered a fully local, browser-based RAG system using WebGPU-optimized embedding models for private hardware acceleration.
  • Integrated Canvas LMS data into a domain-specific knowledge graph to reduce hallucinations in student-facing AI.

Rutgers Department of Computer Science | Research Assistant

Jan 2026 – Present

  • Researching second-order optimizers (e.g., Shampoo) to accelerate convergence in large-scale ML training under Prof. Zhao Zhang.
  • Designing agentic systems for non-sequential code generation, decoupling reasoning from syntax.

Algorithmic Robotics and Control Lab | Research Assistant

Jan 2026 – Present

  • Optimizing robot motion planning algorithms using CUDA for parallel computation (pRRTC and FCIT*).
  • Working with foundation vision-language-action models (NVIDIA GR00T).

Samaritan Scout | Full Stack Engineer

May 2023 – Aug 2025

  • Developed an agentic scraping system using OpenAI/Gemini to autonomously extract data from 100,000+ non-profit websites.
  • Built a React/PostgreSQL search engine indexing 35,000+ volunteer opportunities with semantic search.

Projects

  • Local Coworker (macOS Agent): Native desktop agent using Gemini Vision, macOS Accessibility APIs, and Model Context Protocol (MCP) for autonomous computer use.
  • SLM Optimization (GSM8k): Fine-tuned small language models for math reasoning; deployed on Rutgers High-Performance Computing (Amarel) cluster.
  • HiC-TAD-Library: Python toolkit for 3D chromatin analysis and AI-powered enhancer design; includes a 3D polymer simulation engine using Langevin dynamics.
  • TopCoder Challenge: Ranked 5th globally for reverse-engineering a 60-year travel reimbursement system using XGBoost and feature engineering.
  • NOSTRADAMUS: 3D geospatial intelligence platform for NJ with live flight tracking and layered economic data.

Skills

  • Languages: Python, C++, CUDA, Java, SQL, TypeScript, R, Bash.
  • Frameworks: PyTorch, WebGPU, Transformers, Pandas, Scikit-learn, AWS, HPC Clusters (SLURM).
  • Specialties: Agentic Workflow Design, LLM Post-training, Optimization, Geospatial Intelligence.

Awards

  • Phi Beta Kappa (Inducted Jan 2026)
  • 8090 AI Top Coder (5th Place Worldwide)
  • Rutgers Shark Tank (2nd & 3rd Place)
  • ASA Fall Data Challenge (Top 3)
  • Rutgers Honors College Scholar, Dean’s List.
  • Phi Beta Kappa