Abhilash Bora

Work

A closer look at my work across applied AI engineering, machine learning and research.

AiStrike

Applied AI Engineer · Jan 2025 – Present

I work across LLM applications and the backend systems behind an enterprise security product. My role spans AI workflow design, model evaluation, performance and cost optimisation, production reliability, and technical ownership of customer-facing work.

  • Built and scaled the core multi-agent investigation platform using a custom Python framework, supporting 20+ enterprise customers across multiple production environments.
  • Scaled the processing pipeline to 100K–150K alerts daily, handling peaks exceeding 600K without backlog through improvements to concurrency, queuing and memory management.
  • Led the consolidation of two investigation services into one platform in six weeks, validating the migration against 2.1M production alerts and reducing AI calls by approximately 10× for routine investigations.
  • Designed a deterministic planning path that achieved 99.9% agreement with the AI planner on 10% of alert traffic, allowing eligible investigations to bypass an expensive model call while retaining AI fallback.
  • Improved grouped-alert processing and crash recovery, recording zero alert loss in a 20K-alert fault-injection test. Also led root-cause analysis and fixes for multiple SEV0/SEV1 production incidents.
  • Developed model evaluation and per-customer AI cost tracking to inform model selection and improve the economics of production AI workloads.
  • Owned features from requirements through delivery, working directly with the co-founder/CPO, and lead engineers on customer POCs and investigation features while mentoring interns.

Python, Custom multi-agent framework, OpenAI, Kafka, Redis, PostgreSQL, OpenSearch, Kubernetes, AWS, Langfuse

IIT Guwahati

Summer Research Intern · May 2024 – Jul 2024

Researched improvements to topic representations in BERTopic using variational autoencoders under the guidance of Prof. Ashok Singh Sairam.

  • Explored autoencoder-based parameter estimation for UMAP within the BERTopic pipeline.
  • Achieved a reported 5–10% improvement in topic coherence over the baseline across a corpus of 19,000 documents.

Python, BERTopic, UMAP, Variational Autoencoders, NLP

NUS Singapore

Deep Learning Academic Intern · Dec 2022 – Jan 2023

Worked on deep learning-based sentiment analysis at the NUS School of Computing under Dr. Samantha Sow and Dr. Amirhassan Monajemi.

  • Implemented BiLSTM and CNN models for five-point sentiment classification of Twitter data.
  • Led a five-member team that achieved a 15% improvement over the baseline and received faculty recognition for the work.

Python, PyTorch, BiLSTM, CNN, NLP, Sentiment Analysis

B.Tech, Computer Science and Engineering

Indian Institute of Information Technology, Pune · Jul 2021 – Jul 2025

  • Ranked in the top 1 percentile among 1M+ candidates in JEE Main 2021.
  • Placed among the top 8 teams out of 13,000+ participants in the Google × AMD Solving for India Hackathon 2023.

abhilashbora3107@gmail.com