ML Research Intern in Agentic Runtime Systems

About the role

<div class="content-intro"><div> <h2>Our mission</h2> </div> <div class="bub-content-body"> <div class="field field--name-body field--type-text-with-summary field--label-hidden field__item"> <p>Constructor’s mission is to enable all educational organisations to provide high-quality digital education to 10x people with 10x efficiency.&nbsp;</p> <p>With strong expertise in machine intelligence and data science, Constructor’s all-in-one platform for education and research addresses today’s pressing educational challenges: access inequality, tech clutter, and low engagement of students.</p> <h3><strong>Please send your resume in English only.</strong></h3> </div> </div></div><p><span style="font-size: 14pt;"><strong>Interns are NOT eligible for equipment and other benefits as the position is temporary.</strong></span></p> <p><strong>The Reality</strong></p> <p>The frontier models are converging into commodities; the engine is no longer the differentiator. The scarcity in 2026 is not the model.</p> <p>We are not looking for people to fine-tune Llama on toy datasets or write "better prompts." We are building the runtime systems that turn probabilistic token generation into reliable workers that finish 50-step tasks without spiraling. We are solving the math of multi-step reliability through engineering, not wishful thinking.</p> <p><strong>What You Will Solve</strong></p> <p>You will work on the&nbsp;<strong>Agentic Harness</strong>, treating the LLM as a component within a larger, deterministic control system. Your research will focus on:</p> <ul> <li>Architecting loops that survive error accumulation. You will build verification layers that force the agent to recognize its own hallucinations and self-correct.</li> <li>You will implement systems where the agent compiles its own constraints and the harness enforces them via constrained decoding or external execution.</li> <li>You will optimise KV-cache hit rates and design "restorable compression" strategies to prevent attention decay.</li> </ul> <p><strong>Who We Are Looking For</strong></p> <p>We are looking for&nbsp;<strong>systems thinkers</strong>&nbsp;who happen to know ML, not the other way around.</p> <ul> <li>You know that a demo is not a product. You are interested in the "boring" engineering: state machines, retries, idempotency, serialisation — that actually makes agents work.</li> <li>You are comfortable working with the raw output of the model and building robust software in Python or Rust around it.</li> <li>You understand why dynamic tool loading destroys inference economics. You prefer deterministic guardrails over "prompting it to be nicer."</li> <li>You are willing to read a 500-line trace of a failed agent loop to find the exact moment the context drifted. You care about the "last mile" of reliability.</li> </ul> <h2><span style="text-decoration: underline;"><strong>What We Offer For Interns:</strong></span></h2> <ul> <li><span data-contrast="none">Internship Duration: from 2 months or more</span><span data-ccp-props="{"335559685":720,"335559739":0}">&nbsp;</span></li> <li><span data-contrast="none">Type of internship: </span> <ul> <li><span data-contrast="none">100% remote </span></li> <li><span data-contrast="none">or hybrid</span><span data-contrast="none"> in the office in Bremen (GER), Sofia (BG), Belgrade (SRB), Istanbul (TR), Sakarya (TR), Singapore (SGD)</span><span data-ccp-props="{"335559685":720,"335559739":0}">&nbsp;</span></li> </ul> </li> </ul><div class="content-conclusion"><h2><strong>What We Offer</strong></h2> <ul> <li data-start="199" data-end="1027">💻 Choice of work equipment (e.g., laptop, monitor, etc.)</li> <li data-start="199" data-end="1027">🇬🇧 English classes (iTalki – $130 monthly)</li> <li data-start="199" data-end="1027">⏰ Flexible schedule (we usually work between 09:00/10:00 and 18:00/19:00 CET or EET)</li> <li data-start="199" data-end="1027">👶 Newborn bonus (€500 per child)</li> <li data-start="199" data-end="1027">🧠 Patent r

Skills