AI Research Engineer
Software Engineering, Data Science
Bengaluru, Karnataka, India
About H2Loop
H2LooP.ai is building next-generation sovereign coding models and context platforms for system engineers across embedded systems, robotics, UAV, avionics, automotive, consumer electronics, and industrial IoT.
Our core team comes from Google, Toshiba, Cisco, Bosch, Philips, and NXP.
Hardware-aware software development is the last major frontier where general-purpose AI coding assistants fail consistently and visibly — hallucinated register addresses, MISRA violations, incorrect peripheral initialization sequences, code that compiles but destroys hardware. The root cause is structural: general models are trained on web software; embedded engineering code is sparse, proprietary, and constrained by physical reality.
Our research program is built around closing that gap — through domain-specific training data, models trained with real hardware feedback, and neurosymbolic architectures that combine LLMs with formal methods. You will be a core contributor to that program.
What you'll own
Synthetic data generation
Design, extend, and evaluate our agentic pipeline for generating domain-specific pre-training and instruction-tuning data from public datasheets, reference manuals, and open-source embedded codebases
Invent and test data generation strategies: specification-to-code synthesis, compliance annotation, formal requirement extraction from natural language, multi-step reasoning trace generation from hardware documentation
Evaluate data quality rigorously — not just statistical measures, but whether models trained on the data actually improve on real embedded engineering tasks
Identify the highest-value data gaps in our training corpus and design generation pipelines to fill them
Domain-specific model training and adaptation
Own continued pre-training and instruction-tuning runs across H2Loop's domain-specific model families
Design and evaluate training recipes: data mixture, tokenizer configuration, instruction format, curriculum, and RLHF/RLAIF alignment approaches
Benchmark model families rigorously — not just perplexity, but task-level accuracy on hardware-specific code generation, compliance repair, and specification-grounded reasoning
Maintain H2Loop's model evaluation infrastructure: curated benchmark suites, regression pipelines, and human eval protocols tied to real customer tasks
RL with hardware feedback
Design and run reinforcement learning experiments using real hardware boards as the reward environment — generated code that either works on the hardware or doesn't, producing ground-truth training signal no synthetic benchmark can replicate
Develop reward models and preference data pipelines from hardware pass/fail signals, user feedback, and formal verification outcomes
Investigate and prototype sample-efficient RL approaches suitable for the low-throughput, high-cost signal that physical hardware evaluation provides
Neurosymbolic methods and formal verification
Research and prototype approaches that combine neural code generation with symbolic reasoning and formal analysis tools
Investigate feedback loops between generative models and verification systems — how verification outcomes can improve model behavior over time
Explore training techniques that make model-generated code more amenable to formal analysis without requiring explicit instruction at inference time
Research translation
Monitor the research landscape across the areas relevant to H2Loop's stack: code generation, program synthesis, neurosymbolic AI, continual learning, RL for code, formal verification, and domain adaptation
Run experiments to evaluate whether promising techniques hold up on embedded/systems tasks — many results from general coding benchmarks do not transfer
Produce clear findings that drive product and model decisions: what to adopt, what to discard, and what to invest in further
Required experience
PhD or equivalent research experience in machine learning, NLP, or a closely related field — or 4+ years of industry research with a publication record you can defend
Hands-on experience training or fine-tuning large language models: you have run training jobs, debugged training instabilities, and evaluated results against real task benchmarks, not just held-out loss
Strong foundations in deep learning and the transformer architecture: you understand what is happening during pre-training, instruction tuning, and RLHF, not just how to call the APIs
Rigorous empirical methodology: you design controlled experiments, track what changes between runs, and resist overclaiming from noisy results
Strong Python engineering skills — you can implement ideas cleanly, build evaluation pipelines, and productionize experiments without needing a separate engineering team to translate your notebooks
Strong-to-have
Domain knowledge in formal methods or program verification: familiarity with model checkers (CBMC, Frama-C), theorem provers (Lean 4, Coq, Isabelle), or SMT solvers (Z3)
Experience with reinforcement learning from human feedback (RLHF), AI feedback (RLAIF), or execution-based reward (RL from compiler/test/verifier outcomes)
Knowledge of embedded or systems software: C/C++, RTOS, safety standards (MISRA, AUTOSAR, IEC 61508, DO-178C), hardware abstraction layers, or MCU architecture
Experience with synthetic data generation for language model training — not just data augmentation, but designing generation pipelines that produce novel, high-quality training signal
Published work on code generation, program synthesis, neurosymbolic methods, or domain adaptation for LLMs
Familiarity with the industrial deployment constraints of our customers: air-gapped environments, on-prem inference, compute-constrained hardware
What we don't need
A researcher who optimizes benchmark scores on standard datasets. We are building for a domain where the interesting problems are off the benchmark — hardware-specific, proprietary, constrained by physical reality. We need someone who is motivated by problems that don't have leaderboards yet, who can design their own evaluations, and who cares about the gap between a research result and a system that works reliably in a customer's air-gapped datacenter.
First 90 days
Days 1–30: Get deep on H2Loop's AI stack — our data generation pipelines, model families, RL environments, and verification tooling. Run existing training and evaluation pipelines end-to-end. Form a clear view of where the biggest research leverage is.
Days 30–60: Run a focused experiment: a new data generation strategy, a training recipe improvement, a formal verification repair loop prototype, or an RL reward model evaluation. Produce findings with clear implications for the roadmap.
Days 60–90: Propose a research agenda for the next two quarters. Own at least one research thread end-to-end — from experimental design through evaluation to a concrete product or model outcome.
Compensation & logistics
Competitive early-stage equity + salary
In-person, Bangalore office
Small team — direct access to founders, platform engineers, and applied AI engineers
Support for publishing research where findings are non-proprietary
