I am an AI researcher focused on finding clever, optimized ways to close critical logical gaps—especially when it comes to making advanced models run fast without burning through massive, excess computing resources. I thrive on deep technical problem-solving and love competing with myself to bridge the divide between complex machine learning and low-level systems engineering.
To address alert fatigue and heavy infrastructure overhead in enterprise security, I built a Heterogeneous Graph Transformer (HGT) anomaly detection system in PyTorch Geometric that achieved a 0.977 AUROC and a 0.897 AUPRC. The architecture compressed false positives down to 0.61% while processing full 5-minute log windows in just 4.26 milliseconds on standard consumer hardware. The research is currently in the peer-review submission process.
Beyond that previous work, I am currently researching hardware-software co-design using gem5 to optimize spatial AI workloads, like PointNeXt, on standalone RISC-V pipelines. By profiling execution cycle counts and cache hierarchies during intense matrix loops, this work evaluates how specific hardware modifications can be tailored to efficiently compute complex, real-time AI sensor data.
Right now, I am actively seeking AI/ML engineering or research lab roles in ML, systems, or quantum computing, with software engineering as a strong fallback framework. I am committed to building long-term technical expertise in my field; I am also planning advanced study at ETH Zurich or EPFL, currently targeted for the 2028-29 academic block onward.
Non-negotiables
Question 1: Do I already possess every single skill listed under “Required Qualifications”?
- This protects my commitment to structural clarity by ensuring the target roles where my current mathematical and backend depth serves as the primary foundation, filtering out positions that create immediate operational friction through mismatched prerequisite barriers.
Question 2: Is this a core engineering role (SWE, AI/ML, Systems, or Computer Engineering) focused on heavy backend building?
- This directly honors my focus on hardware-software co-design by ensuring I remain focused on low-level optimization and systems architecture, completely eliminating high-level application abstractions or frontend user-interface layers.
Question 3: Does this job involve writing actual underlying code and architecture, rather than using pre-packaged AI tools or writing prompts?
- This preserves my identity as an autonomous researcher and decision-architect by guaranteeing I am building and scaling the primary computing models from scratch, rather than operating superficially within another platform’s pre-defined constraints.
I maintain a curated pipeline of verified opportunities aligned to my criteria.