Skip to content
View kjgpta's full-sized avatar

Block or report kjgpta

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
kjgpta/README.md
header
typing

PortfoliotracesagePyPIProfile views

LinkedInScholarSubstackEmailResume


👋 About

I'm a Forward Deployed Engineer at TrueFoundry (Bengaluru), helping customers turn LLM platforms into production systems that hold up under real traffic.

Before that I was an ML Engineer at Chubb, shipping GenAI at scale — fine-tuning, agents, RAG, and high-throughput inference.

I also build open-source agent tooling (tracesage) and publish NLP research on evaluation, multilingual systems, and code-switching.

📈 Production impact (Chubb)
SignalResult
LLaMA-3.1 70B + LoRA/QLoRA + RAG+25% accuracy · −15% drift
Multi-agent planner / retriever / verifier+18% factual grounding
vLLM on AKS (A100/H100)−40% p95 · +50% throughput · 10K+/day
HawkHireExplainable AI hiring copilot (internal)

⚡ Featured — tracesage

Local-first observability for LangChain & LangGraph. Drop in one callback — watch every run as a live graph + timeline. No cloud account. No Docker. Featured in Python Weekly #750.

pip install "tracesage[langchain]"
tracesage demo # opens the live UI — no API key needed
tracesage

Docs · GitHub · PyPI · Substack

MCP-aware tool attribution · crash-safe handler · pytest fixture · optional OpenTelemetry export · MIT


🛠 Tech stack

skills

Models & agents PyTorch · Transformers · LoRA/QLoRA · RAG · Multi-agent
LangChain / LangGraph · Hugging Face · OpenTelemetry
Serving & infra vLLM · Kubernetes · AKS · Azure · AWS · Docker
Databricks · CI/CD · MLOps
Research NLP · Code-switching · Eval benchmarks

💼 Experience

RoleOrgWhen
🟢Forward Deployed EngineerTrueFoundryJun 2026 – Present
🔵Machine Learning EngineerChubbJul 2023 – May 2026
🟣NLP Research InternSpeech Lab · NTU Singapore2022 – 2023
B.E. Electrical & ElectronicsBITS Pilani2019 – 2023

📚 Publications

PaperVenue
WhoDunit — culprit detection in mystery storiesarXiv · Dec 2024abs
MALM — mixing-augmented LMs for zero-shot MTAACL-IJCNLP 2022paper
Code-switching LM adaptation (statistical augmentation)ACIIDS 2023paper
Singaporean English–Malay code-switching pointsIALP 2023paper
Data augmentation for transformer AESAISC 2023paper

📖 Google Scholar


✍️ Writing


📊 GitHub stats

statslanguages
streak

🐍 Contribution graph

github contribution snake

🌐 Connect

Full portfolio:kjgpta.github.io
LLM-friendly profile:llms.txt · Markdown

Open to collaborations on LLM tooling, production AI, agent observability, and applied NLP.

EmailPortfolio

footer

Pinned Loading

  1. tracesagetracesagePublic

    Local-first observability for LangChain & LangGraph. Drop in one callback — watch every run as a live graph + timeline.

    Python 27 1

  2. WhoDunIt-Evaluation_benchmark_for_culprit_detection_in_mystery_storiesWhoDunIt-Evaluation_benchmark_for_culprit_detection_in_mystery_storiesPublic

    Benchmark repository for evaluating large language models' deductive reasoning on mystery narratives

    Jupyter Notebook 4

  3. kjgpta.github.iokjgpta.github.ioPublic

    https://kjgpta.github.io/

    HTML

  4. Code-Switch-Language-Modeling-for-English-and-MalayCode-Switch-Language-Modeling-for-English-and-MalayPublic

    Code-Switched Data generation based on Part-of-speech and Language Modeling of the generated text.

    Jupyter Notebook 1

  5. NSC-Code-Switch-AnalysisNSC-Code-Switch-AnalysisPublic

    Code-switching analysis using linguistic techniques like part-of-speech

    Jupyter Notebook

  6. Data-Augmentation-for-Automated-Essay-Scoring-using-Transformer-ModelsData-Augmentation-for-Automated-Essay-Scoring-using-Transformer-ModelsPublic

    A novel data augmentation approach for Automated Essay Scoring and modeling it using Tranformers model

    Jupyter Notebook