Loksai Palyam

Driven by ambition and a determined, independent spirit, I thrive on creatively solving complex challenges as an Associate Engineer. My highly logical and concentrated approach enables me to develop pragmatic and effective solutions, always focused on making a tangible impact.

Experience

Associate Engineer

Virtusa · Sep 2025 – Present

Not yet confirmed

• • Collaborating with clients to understand business use cases, identify opportunities for AI adoption, and explore where AI can deliver meaningful business impact.
• • Building production-grade AI & Agentic AI solutions for enterprise workflows using LLMs, automation, APIs, and structured data.
• • Architected an AI-powered Supplier AP AutoReply system using Google Gemini, Apps Script, and ServiceNow APIs for multilingual inquiry classification, entity extraction, and intelligent P2P ticket routing.
• • Reduced manual AP ticket handling by 70% and saved 160+ operational hours/month through automated payment-status workflows and intelligent ticket processing.
• • Built a GraphRAG-powered Data Lineage Assistant using Python, KùzuDB, Cypher, and LangChain, transforming dbt metadata and SQL dependencies into a knowledge graph.
• • Implemented multi-hop graph traversal for upstream root-cause discovery and downstream impact analysis, grounding LLM responses in actual data dependencies.
• • Engineered multilingual LLM pipelines for classification and entity extraction across Japanese, Mandarin, and English supplier queries.
• • Owned and resolved 10+ critical production incidents across AWS data pipelines, dbt transformations, and enterprise middleware integrations.
• • Working at the intersection of Agentic AI, Knowledge Graphs, GraphRAG, RAG, LangGraph, MCP, and LLM applications to build context-aware and explainable AI systems.

Associate Engineer, Client: Air Liquide

Virtusa · Sep 2025 – Present

Not yet confirmed

• Architected an end-to-end AI-powered Supplier AP AutoReply System using Google Gemini, Apps Script, and ServiceNow APIs, automating multilingual supplier inquiry classification, named-entity extraction, and intelligent P2P ticket routing across APAC affiliates.
• Reduced manual AP ticket handling by 70% and saved 160+ operational hours per month for the GBS APAC P2P team by designing automated real-time payment status workflows.
• Engineered LLM prompt pipelines with context-aware classification and entity extraction logic supporting multilingual supplier queries in Japanese, Mandarin, and English.
• Built an AI-powered GraphRAG assistant for dbt data lineage using Python, KuzuDB, Cypher, and LangChain, transforming dbt metadata and model SQL into an in-memory knowledge graph for explainable lineage analysis.
• Implemented multi-hop graph traversal for downstream impact analysis and upstream root-cause discovery, grounding LLM responses in exact dbt dependencies to eliminate hallucinations.

Technology Intern

Virtusa · Jun 2025 – Aug 2025

Not yet confirmed

• Built a Payload Generator module using Google Gemini API and ReactJS that automated microservice integration test-case generation, cutting developer setup time by 40%.
• Developed NFR analysis modules to auto-surface latency, throughput, and reliability benchmarks from architecture documents using LLM-powered extraction pipelines.

I

Software Engineer Intern

IRMAI · Mar 2025 – Jun 2025

Not yet confirmed

• Reduced analytics UI render time by 95% (10 minutes to under 5 seconds) by migrating analytics workloads from SQLite to ClickHouse with optimized columnar query patterns.
• Deployed scalable analytics microservices on Kubernetes (AKS) using Helm charts and PVC-backed persistent storage, supporting high-availability stateful workloads.
• Built a Neo4j knowledge graph modeling 500+ compliance controls and their relationships, enabling automated gap analysis that replaced a fully manual audit process.
• Delivered a FastAPI control monitoring dashboard with real-time SLA/KPI tracking and GenAI-powered anomaly detection for proactive incident alerting.

Software Engineer Intern

IRMAI.IO · Mar 2025 – Jun 2025

Not yet confirmed

• • Built a Neo4j Knowledge Graph modeling 500+ compliance controls and their relationships, enabling automated compliance gap analysis and reducing manual audit effort.
• • Developed a FastAPI monitoring dashboard with real-time SLA/KPI tracking and GenAI-powered anomaly detection for proactive incident alerting.
• • Improved analytics performance by 95%, reducing UI render time from 10 minutes to under 5 seconds by migrating workloads from SQLite to ClickHouse and optimizing columnar query patterns.
• • Deployed scalable analytics microservices on Kubernetes (AKS) using Helm and persistent storage for high-availability workloads.
• • Automated cloud infrastructure provisioning across AWS and Azure using Terraform, reducing manual provisioning overhead for engineering teams.

Academic Hub Chair

Internet Society · Oct 2023 – May 2025

Not yet confirmed

• SRIT Academic Hub — Internet Society Hyderabad Student Chapter
• • Led the Academic Hub and organized technical seminars, workshops, bootcamps, and learning sessions for 600+ students.
• • Conducted technical sessions on topics including Git, Cybersecurity, Cloud Computing, Backend Engineering, and AI.
• • Delivered seminars and workshops focused on helping students understand emerging technologies through practical demonstrations and hands-on learning.
• • Participated in and contributed to technical bootcamps, including DDCN, while coordinating learning activities for students.
• • Led technical discussions and knowledge-sharing sessions, developing strong public speaking, technical communication, and developer community-building skills.
• • Collaborated with students and technical communities to identify learning needs and organize relevant technical sessions and workshops.

Project Intern

Wipro · Aug 2023 – Nov 2023

Not yet confirmed

• • Built a multilingual translation system by fine-tuning M2M100 Transformer models across 12 language pairs, achieving BLEU 0.35 and ROUGE-L 0.42 on held-out evaluation datasets.
• • Designed a Human-in-the-Loop feedback pipeline to identify low-confidence translations for expert review, reducing downstream translation errors by 15%.
• • Worked on NLP, Transformer models, multilingual AI, model evaluation, and human feedback workflows.

What drives their work

The Inventor

You conceive and construct novel technical systems.

Loksai's superpower

You excel at conceptualizing and implementing unique, scalable technical solutions that address complex problems.

How Loksai works

You would excel in a team focused on developing novel AI/ML applications and scalable data infrastructure.