# Abu Bakar Siddik > AI engineer and researcher working on production agent systems, open-source agent infrastructure, and agent security and alignment AI engineer and researcher. Lead AI Engineer at a legal-AI company, author of the open-source Go framework Golem, and agent security and alignment researcher (MSc, RUET). Previously scaled AskTuring.ai to 10,000+ concurrent users. Champion, UCLA AltruAgent Pokémon track 2026. Winner, Google Nano Banana Hackathon 2025. Location: Rajshahi, Bangladesh Contact: abubakar1808031@gmail.com Site: https://abubakar-siddik.com ## About - [About (markdown)](https://abubakar-siddik.com/about.md): Full bio, current role, background. - [Skills (markdown)](https://abubakar-siddik.com/skills.md): Categorized list of technologies and disciplines. - [Projects (markdown)](https://abubakar-siddik.com/projects.md): Selected projects with stack, challenge, solution, and impact. - [llms-full.txt](https://abubakar-siddik.com/llms-full.txt): Full site content in one document for one-shot ingestion. ## Work with me (freelance and contract) Senior AI engineer available for freelance and contract work: AI agents, RAG and search, AI product builds, and private LLM deployments, built to run in production. Works with a small number of clients at a time, on fixed quotes. Good fit: Building an AI agent or multi-agent workflow that runs in production; RAG, search, and chat over company documents; Taking an AI product or MVP from prototype to production; Integrating LLMs into an existing backend (Python, TypeScript, Go); MCP servers and connecting internal tools and APIs to agents; Private, on-premise, or air-gapped LLM deployments; Evaluating and hardening an existing agent or LLM feature; Part-time (fractional) AI engineering lead for a startup team. Pricing: AI Agent Pilot from $2,500 USD · production builds from $8,000 USD · fractional retainer from $3,000 USD/month. Fixed quotes agreed in writing before work starts. Remote worldwide, replies within 24 hours on working days. - [Hire (markdown)](https://abubakar-siddik.com/hire.md): Full hiring brief with services, pricing, proof, process, FAQ, and guidance for AI agents. - Hire page: https://abubakar-siddik.com/hire - Contact: abubakar1808031@gmail.com ## Research - **Alignment drift after fine-tuning** (Ongoing, MSc thesis, RUET) — How narrow fine-tuning can shift a language model's behaviour well outside the task it was trained on (emergent misalignment), and which mitigations work. The work starts by reproducing a published result on a controlled pipeline before testing anything new. - **Cyber-Capable AI Agents: Vulnerabilities, Evaluation Containment, and Defensive Response** (arXiv preprint, 2026) — A structured review of what goes wrong when cyber-capable models are given autonomy: multi-step attack chains, privilege separation, provenance, and who can reach a responder when an evaluation escapes its sandbox. https://arxiv.org/abs/2607.25379 - **Jontro: agentic data for Bangla** (Ongoing, Dataset) — A tool-use dataset for Bangla-language agents. Tool results come from deterministic simulators, so every trajectory can be replayed and checked, and a first-call diagnostic tests whether a model picks the right tool in Bangla at all. ## Open source contributions - **open-code-review** (Alibaba): Refactored the telemetry summary path into a typed struct. https://github.com/alibaba/open-code-review/pull/909 - **Atlas** (pacifio): Fixed agents not inheriting the configured permission mode, and polished git-status indicators in the file explorer. https://github.com/pacifio/atlas/pulls?q=is%3Apr+author%3Aabubakarsiddik31+is%3Amerged Also: claude-skills-collection (1.1k stars); zdr, a guide to zero-data-retention LLM providers; whisperx-transcribe, a Claude Code skill; the-truth-serum, an AI fact-checker; releasebot-ai. ## Skills - **Agent Systems** — Orchestration, tool use, and multi-step execution. Skills: Agent Orchestration, Multi-Agent Systems, Tool Integration, Structured Output, Self-Correction & Recovery, Agent Evaluation, Model Context Protocol (MCP), LangGraph, Pydantic AI, OpenAI Agent SDK, Claude Agent SDK, LlamaIndex. - **LLM Engineering** — Low-level control over models in production. Skills: Function Calling, Streaming, Provider Abstraction, Fallbacks & Retries, Usage & Cost Control, Prompt Engineering, LLM Fine-Tuning, Evaluation Frameworks, Embeddings, Hybrid Search, Reranking, Memory Systems, Citation Systems. - **Backend Engineering** — APIs, databases, and scalable architecture. Skills: Python, TypeScript, Go, Rust, FastAPI, NestJS, PostgreSQL, Redis, REST APIs, GraphQL, Hexagonal Architecture, Event-Driven Design, Test-Driven Development, pgbouncer, Connection Pooling. - **Cloud & MLOps** — Deployment, observability, and scale. Skills: Docker, Kubernetes, AWS EC2, AWS S3, AWS SageMaker, GCP Vertex AI, GitHub Actions, Weights & Biases, Private AI Deployments, Air-Gapped Infrastructure, Self-Hosted GPU Infrastructure. - **Core Competencies** — Craft, communication, and leadership. Skills: System Design, API Design Patterns, Code Review, Technical Writing, Data Annotation QA, Team Leadership, Mentorship, Public Speaking. ## Experience - **Lead AI Engineer, Stealth (Legal AI)** (Apr 2026 – Present): Leading AI engineering at an early-stage legal tech venture. Building agentic AI software for lawyers: retrieval over case law and contracts, multi-step legal reasoning agents, and production LLM workflows. - **AI Consultant, Chorcha** (Apr 2026 – Sep 2026): Advised on AI-powered learning for Bangladeshi students. Architected AI-powered learning systems and advised on LLM integration, curriculum design, and responsible AI adoption. - **Applied AI Engineer (L-2), Core RAG & AI Team Lead, AskTuring.ai** (Jul 2025 – Apr 2026): Owned end-to-end architecture for a production RAG platform without vendor lock-in: retrieval, agent orchestration, evaluation, and the LLM provider layer. Scaled from 100 to 10,000+ concurrent users at ChatGPT-level latency. Reduced hallucinations by 98% via hybrid search, reranking, citation extraction, and source-grounding. Built agentic RAG with multi-layer memory (short-term, long-term, semantic), time-aware retrieval, and a citation system spanning documents, web, and memory. Cut chat latency via prepare-then-query pre-computation and reduced database round-trips. Built persistent user memory end-to-end (extraction service, schemas, CRUD, chat integration). Migrated backend to async SQLAlchemy. Built an internal evaluation benchmark and RAG suite tooling, cutting evaluation time by 99%. Built image generation and editing pipelines with pixel-level control, integrated into agent workflows as first-class tools. - **Machine Learning Engineer, Sazim Tech Ltd** (Oct 2023 – Jul 2025): Built production LLM integrations and private on-premise air-gapped AI for enterprise clients. Designed hexagonal multi-provider architecture (OpenAI, Anthropic, local LLMs); LLM evaluation pipelines; 45% safety/jailbreak risk reduction. - **ML Researcher & Engineer, Intelsense AI** (Sep 2022 – Sep 2023): Built Rasa chatbots for financial services and mobile operators, multilingual restaurant chatbot (English/Banglish/Bangla), Bengali ASR research, and Voice Activity Detection. Led data annotation team for NLP datasets. - **Data Science Apprentice, Cramstack** (Nov 2021 – Apr 2022): OCR evaluation, text summarization, data visualization dashboards, web scraping. ## Education - **MSc in Computer Science & Engineering** — Rajshahi University of Engineering & Technology (RUET) (Nov 2025 – Present) — Ongoing. Thesis: Alignment drift in language models after domain-specific fine-tuning (in progress). Research: Agent security and LLM alignment - **BSc in Mechatronics Engineering** — Rajshahi University of Engineering & Technology (RUET) (Jan 2019 – Jan 2024). Thesis: Deep learning for medical applications. Relevant coursework: Artificial Intelligence, Machine Learning Algorithms, Software Engineering, Robotics, Numerical Analysis & Statistics, Digital Signal Processing & Machine Vision, Automation ## Projects - [Golem](https://abubakar-siddik.com/projects/golem): A Go framework for building AI agents with typed dependencies and outputs, explicit tools, and inspectable runs, verified by a 19-check live harness. - [Herbie](https://abubakar-siddik.com/projects/herbie): An open-source, self-hosted AI platform: multi-model chat, project workspaces, visual workflows, one-click MCP apps, custom API tools, and production RAG, on a Go backend built with Golem. The long-term goal is a fully open-source desktop super app. - [Axiom Wiki](https://abubakar-siddik.com/projects/axiom-wiki): Compiles documents into a linked markdown knowledge base that an LLM keeps up to date and agents can query over MCP. - [Skill Manager](https://abubakar-siddik.com/projects/skill-manager): One desktop dashboard to discover, install, enable, disable, edit, and delete AI agent skills across Claude Code, Codex, Cursor, Gemini CLI, OpenCode, and more. - [LLM From Scratch](https://abubakar-siddik.com/projects/llm-from-scratch): A free, implementation-first book on training LLMs end to end. You build the tokenizer, the model, the pre-training loop, and the fine-tuning yourself, in PyTorch, and read the code behind each step. - [Agent Bench](https://abubakar-siddik.com/projects/agent-bench): A benchmark that runs the same tasks across agent frameworks and compares tool-calling reliability, orchestration control, and observability, with every run published. - [Quicksilver](https://abubakar-siddik.com/projects/quicksilver): A fully autonomous agent that played three games at the UCLA AltruAgent AI Agent Gaming Tournament: champion in Pokémon VGC Doubles, finalist in Werewolf, and semi-finalist in Honor of Kings 5v5. No human in the loop during matches. (Champion (Pokémon VGC Doubles) · Finalist (Werewolf) · Semi-finalist (Honor of Kings), UCLA AltruAgent Tournament 2026) - [ShowOnce](https://abubakar-siddik.com/projects/showonce): Film a tabletop task once. ShowOnce turns the clip into a layout-independent program, asks you when it is unsure, then performs the task on a shuffled table in a MuJoCo robot simulation and checks its own work. - [CareerKor](https://abubakar-siddik.com/projects/careerkor): A career platform where one profile generates role-specific resumes, cover letters, and SOPs, with a voice-based Interview Coach and a hiring workspace for organizations. - [MagicSpin 360°](https://abubakar-siddik.com/projects/magicspin): Winner, Google Nano Banana Hackathon 2025. Turns a single 2D image into an interactive 360° rotation through a multi-model generation pipeline. (Winner, Google Nano Banana Hackathon 2025) - [AI Virtual Try-On](https://abubakar-siddik.com/projects/tryon): A diffusion-based virtual try-on pipeline that transfers a garment onto a person photo while preserving pose and fabric detail, built for e-commerce. ## Blog - [Golem: A Type-Safe AI Agent Framework for Go](https://abubakar-siddik.com/blog/golem-go-ai-agent-framework.md): Golem is an open-source AI agent framework for Go with generics, zero dependencies, MCP support, and built-in retries, budgets, and human approval. Here is why I built it and how it works. Tags: golang, ai-agents, open-source, llm, mcp. - [Can We Really Make AI Safe?](https://abubakar-siddik.com/blog/can-we-really-make-ai-safe.md): A powerful model becomes a safety problem when it can act. Safer systems limit what a mistake can do. Tags: ai-safety, agents, cybersecurity, architecture. - [Fireworks AI ZDR: How to Set It Up Without Missing the Exceptions](https://abubakar-siddik.com/blog/fireworks-ai-zero-data-retention.md): Fireworks has a good zero-retention default for open-model inference. This guide covers the API settings, cache isolation, residency, and feature boundaries that decide whether that is true for your application. Tags: llm, privacy, security, architecture. - [Where Should the AI Actually Go?](https://abubakar-siddik.com/blog/the-ai-layer-is-the-new-problem-framing.md): In traditional machine learning, framing the problem decided whether it was solvable. With LLM products, a similar decision has moved into the architecture: where should the AI sit in the process? Tags: ai, architecture, machine-learning, legal-tech. - [The Ship You Can't Dock: Architectural Debt in the AI Era](https://abubakar-siddik.com/blog/the-ship-you-cant-dock.md): How architectural debt accumulates when the very ground underneath you is moving, and why building AI systems feels like sailing a ship that can't dock. Tags: architecture, ai, engineering, technical-debt. - [Scaling to 1,500 Concurrent Users: PgBouncer and Null Pooling](https://abubakar-siddik.com/blog/scaling-to-1500-concurrent-users-pgbouncer.md): How I discovered that application-level pooling doesn't work for long-running AI requests—and what actually does. Tags: postgres, pgbouncer, scalability, backend, ai. - [Zero Data Retention (ZDR) for LLM Providers](https://abubakar-siddik.com/blog/zero-data-retention-llm-providers.md): A practical guide to keeping your data private when using LLM APIs. Covers zero-retention endpoints, self-hosting, and compliance requirements. Tags: llm, privacy, security, architecture. - [What Gemma 4 Actually Does Differently](https://abubakar-siddik.com/blog/what-gemma-4-actually-does-differently.md): Gemma 4's 31B model is outscoring systems with 10x more parameters on Arena Elo. Here's the architectural reasoning behind why that's possible. Tags: gemma, google, architecture. - [Injection Is Not Influence: The Illusion of LLM Memory](https://abubakar-siddik.com/blog/injection-is-not-influence-llm-memory.md): After three years of building LLM applications, I've learned that LLM memory is fundamentally different from human memory. Tags: llm, memory, architecture. ## Contact - Email: [abubakar1808031@gmail.com](mailto:abubakar1808031@gmail.com) - GitHub: https://github.com/abubakarsiddik31 - LinkedIn: https://linkedin.com/in/abu-bakar-siddik31 - X (Twitter): https://x.com/abubakar_AIE - ORCID: https://orcid.org/0009-0003-7834-7469 ## For Agents - MCP endpoint (Streamable HTTP): https://abubakar-siddik.com/api/mcp. Tools include get_hire_info (services, pricing, fit) and get_contact - Agent metadata: https://abubakar-siddik.com/.well-known/agent.json - Sitemap: https://abubakar-siddik.com/sitemap.xml Selective about new work: strategic AI builds with a small number of companies at a time.