Skip to content
View Ikteder's full-sized avatar

Achievements

Achievement: Pull Shark

Achievements

Achievement: Pull Shark

Block or report Ikteder

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Ikteder/README.md

Ikteder Akhand Udoy

PhD researcher in Computing at Boise State University, working on ML systems, efficient AI, LLM inference, quantization, and computer vision. Expected graduation: December 2027. Previously Software Engineer I at Playense.

Seeking Summer 2027 internships in machine learning engineering, applied AI research, ML systems, data science, and software engineering.

Portfolio & résumé · LinkedIn · Google Scholar · Email

I build reproducible Python/PyTorch experiments, inference benchmarks, deployment checks, and tools that make evidence easy to inspect.

Selected work

Project Focus
Po2QAT Reproducible power-of-two quantization-aware training across compact CNN, ViT, and language-model examples
SCALE-Sim for SPARQ Scheduling and memory-traffic simulation for constrained weighted subgraphs across five compact models; not measured silicon or end-to-end speedups
LLM Agent Workflows Grounded experiment-report retrieval, run comparison, tool routing, and cited summaries
Model Export & Benchmarking Model export, cross-runtime parity checks, inference benchmarks, and deployment reports
SchemaRoute Offline JSON Schema preflight against dated LLM provider profiles
StreamReceipt Completion-evidence checks for captured LLM tool-call streams
AgentTrace Canary Offline structural regression checks for exported agent traces
RankQuake Sensitivity and uncertainty analysis for weighted multi-criteria decisions

Research and engineering interests

  • Efficient and hardware-aware machine learning
  • Quantization, deployment verification, and model evaluation
  • Computer vision and trustworthy dataset tooling
  • LLM reliability and structured-output testing
  • Native mobile, systems, networking, and developer tools

Working principles

  • Tie claims to tests, metrics, datasets, logs, or reproducible commands.
  • Document dataset provenance, model limitations, and experiment conditions.
  • Prefer deterministic demos and local-first tools when they make evaluation easier.
  • Treat accessibility, safe defaults, and honest failure modes as product requirements.

Technologies

Python · PyTorch · scikit-learn · TypeScript · React · Node.js · Kotlin · Android · C++ · FastAPI · SQL · GitHub Actions

More

Pinned Loading

  1. Customer-Churn-Revenue-Risk-Analytics Customer-Churn-Revenue-Risk-Analytics Public

    Customer churn and revenue-risk analytics project with reproducible modeling and business-facing outputs.

    Python

  2. Demand-Forecasting-and-Inventory-Decision-Support Demand-Forecasting-and-Inventory-Decision-Support Public

    Reproducible retail-demand forecasting demo with chronological evaluation and inventory-risk summaries.

    Python

  3. predictive-maintenance-showcase predictive-maintenance-showcase Public

    Predictive-maintenance showcase for early-warning analysis, risk visualization, and reproducible demos.

    Python

  4. dataset-quality-leakage-detection-platform dataset-quality-leakage-detection-platform Public

    Pre-training quality gate for dataset defects, duplicates, split leakage, and label problems.

    Python

  5. ml-experiment-control-center ml-experiment-control-center Public

    FastAPI and React control center for launching, tracking, comparing, and exporting ML experiments.

    Python

  6. Po2QAT-Power-of-Two-Quantization-Algorithm Po2QAT-Power-of-Two-Quantization-Algorithm Public

    Reproducible power-of-two quantization-aware training examples for compact CNN, ViT, and language models.

    Python 11