Skip to content
View Buljic's full-sized avatar
👋
👋
6 followers · 2 following

Achievements

Achievement: YOLOAchievement: Pull Sharkx2

Achievements

Achievement: YOLOAchievement: Pull Sharkx2

Highlights

  • Pro

Block or report Buljic

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
Buljic/README.md

Imad Buljić

Software Engineer • Team Lead • Researcher • MA Student

Building software systems, leading engineering teams, and publishing research across different fields.

Website • LinkedIn

Animated banner

What I am building now

Two projects I would point an engineer to first:

  • llm-serving-lab - a measurement harness for LLM serving. Seeded Poisson traffic, streaming TTFT and TPOT capture, goodput against a stated SLO, cost per million output tokens. Run for real against a live server, raw rows committed, every table regenerated from them.
  • agent-workbench - an internal platform for tool-calling agents: path and command guardrails, a write approval gate, a per-step audit trail, and an offline eval harness whose tests fail when a guard is removed.

About

I am a software engineer and a team lead

Alongside industry work, I am an MSc student at the University of Zenica, researcher and a BHFF Scholar. Outside all of that, I do like 400 random things and also post photography stuff on @imads.art.

Online

Pinned Loading

  1. Java_Transactional_Outbox Java_Transactional_Outbox Public

    Exploring the transactional outbox pattern in Java: save business state and an outbox event in one PostgreSQL transaction, then dispatch to Kafka with retries and idempotency.

    Java

  2. SLO_Preventive_Autoscaling_KEDA SLO_Preventive_Autoscaling_KEDA Public

    Reproducible demo of SLO-aware event-driven autoscaling with Go, NATS JetStream, Prometheus, Grafana and KEDA on kind.

    TypeScript

  3. deepseek-v4-customMiniVer deepseek-v4-customMiniVer Public

    A local code-agent lab: structural repo ingestion, verifier-guided context retention, and benchmark tooling.

    Python 1

  4. medlineplus-rag medlineplus-rag Public

    CPU-only RAG system over MedlinePlus health topics: hand-written hybrid retrieval with exact-span citations and a reproducible evaluation harness

    Python

  5. agent-workbench agent-workbench Public

    Internal AI enablement platform: tool-calling agents against a repository with path and command guardrails, an approval gate, a per-step audit trail, and an offline eval harness with a guard mutati…

    Python

  6. llm-serving-lab llm-serving-lab Public

    Bench harness for LLM serving: seeded Poisson workloads, streaming TTFT/TPOT capture, goodput vs a stated SLO, cost per million tokens. Ships a real llama3.2 CPU campaign and a GPU sweep config.

    Python