We help eng teams catch critical blind spots and boost engineering excellence at every commit.

Making AI agents find impactful and tested edge cases for PRs, reliably.

Engineer #1. Code change analysis, codebase search, navigation & analysis, test suite generation and maintenance workflows, testing infra (repo configuration, sandbox orchestration), among other things.

Developed and productionized high-performance and reliable AI systems to deliver high-impact features, enabling automated two-way email conversations in 6sense Conversational Email:

• Hyper-personalized AI outreach emails

• AI-powered reply automation

𝗥𝗲𝗽𝗹𝘆 𝗔𝘂𝘁𝗼𝗺𝗮𝘁𝗶𝗼𝗻: Fully automating >70% of live lead emails with 99.9% accuracy in production, achieving the best outcomes in the tradeoffs between accuracy, latency and human effort for scalability.

• Text classification, model ensembles, NER, outlier detection (Transformers, PyTorch)

• Human-in-the-loop ML, dispatch/validation algorithms, active learning, automatic consensus-based ground truths

• MLOps (streaming, serverless inference, data pipelines, ONNX/quantization)

• Reliable ML (data quality, testing, monitoring and observability, analytics)

𝗔𝗜 𝗘𝗻𝗴𝗶𝗻𝗲𝗲𝗿𝗶𝗻𝗴: end-to-end AI systems with RAG, prompt engineering, and bespoke workflows; spearheaded and deployed initiatives for ensuring reliability of LLM outputs

• Semantic search backend service (FastAPI, Pinecone/pgvector, Sentence-Transformers)

• LLM online evaluation / guardrails for trustworthy, reliable, and appropriate LLM outputs

• Steerable LLM generation (custom intents and reply strategies)

Research project:

Generalized acoustic analytics for scalable novelty detection in smart manufacturing

(in other words, predictive maintenance using sound data from manufacturing equipment)