GIHAN LAKMAL
AI/ML EngineerAI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
AI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
AI/ML Engineer with 2+ years in Machine Learning, Generative AI & Multi-Agent Systems
AI/ML Engineer with nearly 2 years of experience developing, deploying, and optimizing machine learning, deep learning, and generative AI solutions. Skilled in Python, data preprocessing, feature engineering, model evaluation, time-series forecasting, NLP, computer vision, RAG, LLM applications, and multi-agent AI systems. Experienced with FastAPI, LangChain, LangGraph, ChromaDB, Neo4j, PostgreSQL, Docker, MLOps, DevOps, CI/CD, monitoring, logging, model tracking, and performance optimization for scalable AI solutions and business-focused stakeholder insights across real-world production projects and deployments.
University of Kelaniya - Sri Lanka
B. Sc. (Hons) Computer Science (AI Specialization) · Computer Science (AI Specialization)
August 1, 2022 – June 30, 2026
University of Colombo - Sri Lanka
Bachelor of Information Technology (External) · Information Technology
August 1, 2022 – June 30, 2023
Nalanda College, Colombo 10
G.C.E. Advanced Level 2020 Physical Science Stream · Mathematics, Physics, Chemistry
June 1, 2020 – May 31, 2020
ArcheAl
Co Founder
December 1, 2025 – Present
India
SenzMate AIOT
AI/ML trainee Associate Engineer
July 1, 2025 – Present
India
FinTech MultiAgent System For StockBroker Firm - Vinvest Client Project
June 1, 2026 – June 1, 2026
Built AI platform that combines multi-agent LLM orchestration (LangChain/LangGraph), vector search (ChromaDB), and a Neo4j knowledge graph to provide retrieval-augmented conversational analytics and automated reporting. Built an AI-powered analytics platform combining multi-agent LLM orchestration, RAG, vector search, and Neo4j knowledge graphs to deliver conversational insights and automated reporting. Designed and implemented scalable backend pipelines using FastAPI, LangChain/LangGraph, ChromaDB, Neo4j, and Docker, including document ingestion, token-aware chunking, embeddings, metadata enrichment, and hybrid retrieval. Demonstrated capability in building production-ready AI systems with agent workflows, graph-based reasoning, observability, async/background processing, cost optimization, and reliable deployment practices. Tech Stack: Python, FastAPI, LangChain, LangGraph, ChromaDB, Neo4j, Cypher, RAG, PDF/Word/Excel ingestion, token-aware chunking, NER metadata enrichment, embeddings, tiktoken, LangSmith, Docker, Docker Compose, async background workers, caching, structured logging, health checks, correlation IDs, Agent Graph, Deep agents, Agent orchestration...
Medical Drug Sales Prediction based on Specific areas in Sri Lanka (IEEE Research Publication)
June 1, 2026 – June 1, 2026
Built production-grade time-series forecasting and inference pipeline for weekly, area-level pharmaceutical sales. Modeling: Architected and ensembled GRU/LSTM, Transformer/TFT, N-BEATS, LightGBM/XGBoost, and Prophet models with time-series-aware cross-validation. Feature Engineering: Implemented lag/rolling features, calendar/holiday encodings, seasonal decomposition, imputation, and scaling pipelines for robust sequence learning. Explainability & Causality: Delivered SHAP-based local/global explanations and causal discovery/counterfactual analyses to surface actionable drivers. Deployment & MLOps: Containerized inference/API (Docker, Flask/FastAPI), CI tests, experiment tracking, and monitoring for model drift and reproducibility.
View ProjectAdvanced Agentic Hybrid OCR PROJECT - Client Project ArcheAl
June 1, 2026 – June 1, 2026
Designed and deployed a production-grade OCR and AI document processing system for structured data extraction from complex multi-page insurance approval documents, using Tesseract OCR, OpenAI Vision/VLM models, validation rules, and AI-based gap filling to achieve 94-98% extraction accuracy. Built a 7-phase asynchronous processing pipeline with Temporal, FastAPI, PostgreSQL, SQLAlchemy, and Docker, including job tracking, extraction storage, audit logging, multi-tenant support, security controls, and optimized throughput of 100+ documents per hour. Tech Stack: Python, FastAPI, Temporal, PostgreSQL, OpenAI Vision Models, Tesseract OCR, React, Docker, SQLAlchemy, RAG system, Agent Orchestration, LLM Ochastration....
View ProjectMedical Drug Sales Prediction based on Specific areas in Sri Lanka (IEEE Research Publication)
IEEE
June 1, 2026 – Present
Cultural Fit Analysis
The candidate's experience as a Co-Founder at an AI research and engineering startup, coupled with an AI/ML trainee role, demonstrates a proactive and innovative mindset. The diversity of projects (FinTech, Medical Sales Prediction, OCR) indicates adaptability and a broad interest in applying AI across different domains. The focus on production-grade systems and MLOps suggests a practical, results-oriented approach, which aligns well with a dynamic, fast-paced technical culture. The IEEE research publication also points to a commitment to continuous learning and contributing to the field.
Soft Skills & Operational Fit
The candidate's project descriptions highlight end-to-end ownership, problem-solving, and the ability to work on real-world client projects, indicating strong operational fit. Experience as a Co-Founder suggests leadership, initiative, and a product-oriented mindset. The detailed descriptions of complex AI systems imply strong analytical and problem-solving skills. However, without direct assessment data on communication or teamwork, these are inferred from the quality of project descriptions.