Portfolio
LinkedIn · GitHub · niranjankrishna.acad@gmail.com
Experience
Application Engineer, Formant (Nov 2022 – Apr 2025)
- Led development of the core fleet management portal, contributing to 20–30% of company revenue, with 99.8% uptime for 20,000+ production robots.
- Engineered core libraries and pipelines (Python, Go, gRPC, ROS2) for real-time robot control and communication, improving CSAT to >92%.
- Built fault-tolerant data ingestion pipelines, processing 500k+ records per month and reducing latency from 15 minutes to sub-second.
- Developed performance analytics with Snowflake and Python, handling 1M+ data points in real time for operational insights.
AI Engineer, Reknow.ai (Nov 2021 – Apr 2022)
- Fine-tuned GPT-J locally for chatbot applications, saving over $200k in inference costs.
- Designed clustering-based QA models in PyTorch, achieving ~85% accurate query resolution.
Lead Software Engineer, FindMonster (Jan 2021 – Jul 2021)
- Built AR gameplay with Niantic ARKit, incorporating real-world environment awareness.
- Implemented semantic segmentation models for natural-object placement with IoU >0.7.
Lead Software Engineer, TheGGLife (Jan 2020 – Dec 2020)
- Engineered Node.js server architecture for live streaming, supporting 150k+ concurrent viewers via WebSockets.
- Developed Unity multiplayer games with live-stream integration and PyTorch NLP models for command-to-gameplay translation (~94% accuracy).
Projects
LedPulse — AI-Powered Computational Art Installation (Dec 2024)
- Built a Python pipeline for Dragon (volumetric LED display) that transforms live speech into abstract musical visuals via small language models analyzing emotion and tone.
Publications
- “Classier Guided Diffusion for Image Inpainting: Applications to Fine Art.” Accepted at LXAI @ ICML 2022.