AI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit

Sr SDE @ Amazon | LLM Inference Engineer — Driving SOTA Latency at Scale | Speculative Decoding · CUDA · Distributed ML Systems | undertheinferencehood.substack.com
I solve complex, ambiguous problems at the intersection of AI models, hardware, and distributed systems — and ship solutions that serve millions of customers. As a Senior SDE at Amazon, I lead inference optimizations for the Amazon Nova family of models — spanning GPU and custom silicon, from single-node to multi-node frontier scale. LLM inference is where research and production speak different languages. A technique that shows gains in a paper often falls apart against real hardware constraints, feature cross-compatibility (multimodal, customization, constrained decoding), quantization trade-offs, and the tension of keeping accelerators maximally utilized while serving low-latency responses. Add external dependencies with uncertain timelines and multiple teams with conflicting hypotheses — and the right approach often doesn't exist yet. That's where I thrive: → Led speculative decoding across Amazon Nova models — significant throughput improvements with multimodal, constrained decoding, and LoRA cross-compatibility → Unblocked a frontier-scale model launch by debugging accuracy regressions and delivering major throughput gains under multi-node constraints → Implemented speculative decoding for LoRA, enabling low-latency model customization for a critical product launch → Root-caused a critical latency regression, then delivered substantial performance improvements → Wrote custom CUDA kernels for performance-critical inference paths → Filed 3 patents in inference optimization How I work: → End-to-end ownership from ambiguous requirements through production deployment → Bias for action — I make decisions with incomplete data, move fast, and course-correct → Cross-team alignment across model training, evals, runtime orchestration, and external partners → Force multiplier through mentoring, knowledge sharing, and scalable frameworks → T
Gayatri Vidya Parishad College of Engineering (Autonomous)
Bachelor's Degree, Computer Science
N/A – Present
Amazon
Senior Software Development Engineer
April 1, 2021 – Present
Amazon
SDE-II, Alexa ML Data Platform at Amazon
April 1, 2016 – April 1, 2021
Amazon
SDET-II
December 1, 2013 – April 1, 2016
Amazon
SDET - II
April 1, 2013 – December 1, 2013
Amazon
Software Developer Engineer in Test
July 1, 2011 – April 1, 2013
Akamai Technologies
Software Engineer
September 1, 2009 – July 1, 2011
Bangalore
Yahoo!
Quality Engineer
June 1, 2008 – September 1, 2009
Cultural Fit Analysis
The candidate has a long tenure at Amazon, indicating stability and experience within a large, fast-paced tech environment. The progression from SDET to Senior SDE demonstrates adaptability and growth. However, the lack of diverse company experience outside of Amazon (since 2011) and Akamai/Yahoo prior to that, combined with no listed personal projects, limits the assessment of broader cultural fit and adaptability to different organizational structures or startup environments. The target role of ML Engineer aligns well with their recent experience in LLM inference.
Soft Skills & Operational Fit
The candidate's experience at Amazon, particularly in leadership and cross-team collaboration, suggests strong operational fit and soft skills relevant to a senior role. Mentoring and unblocking critical launches indicate problem-solving and leadership capabilities. However, without specific psychometric or communication test results, a detailed assessment of soft skills like stress handling or team collaboration is limited.