.png&w=256&q=75)
Ananya Raj@ananyarajj
4
Events attended
2
Submissions made
India
1 year of experience
About me
Passionate about building AI systems that solve real problems, not just impressive on paper but actually useful in the real world. My work spans machine learning, NLP, and full stack AI development, from designing ensemble classifiers to deploying intelligent systems that non-technical users can actually interact with. I enjoy working at the intersection of research and product, where a model isn't done until it's wrapped in something people can use. Driven by curiosity, obsessed with finding gaps nobody else noticed, and genuinely excited about what becomes possible when AI is used with intention.
🤝 Top Collaborators
.png&w=640&q=75)
🤓 Latest Submissions
Sahayak: Safety Aware Voice Interpretation Agent
A doctor asks, “How often are you taking this medicine?” The patient knows the answer, but not in the same language. In healthcare, that gap can matter. Sahayak is a real-time English-Hindi clinical voice interpretation platform built to make conversations between doctors and patients clearer, safer, and easier to document. Unlike a regular translation tool, Sahayak is designed around the things that cannot afford to get lost in translation: medicine names, dosages, symptoms, allergies, numbers, and urgent warning signs. During a consultation, both sides can speak naturally while Sahayak provides translated audio and text, checks clinically important information, flags uncertain phrases, and highlights emergencies that may need immediate attention. Doctors and patients can also verify extracted medical facts in their own language instead of simply trusting a translated sentence. Once the conversation ends, Sahayak turns it into a structured bilingual consultation record with timestamps and verified clinical information. The idea is simple: patients should not need perfect English to explain what hurts, and doctors should not have to guess what their patients mean.
30 Sep 2026
EchoSync - From Conversation to Code Autonomously
EchoSync is a fully autonomous meeting-to-GitHub pipeline for engineering teams. Give it an audio file or transcript and it handles the rest: transcription, task extraction, ownership assignment, story point estimation, contradiction detection, and GitHub issue creation — automatically. The four-layer assignment engine determines task ownership using regex-based verbal detection, LLM contextual inference, simulated Git blame, and team roster expertise matching as a final fallback. The Semantic Memory Vault stores architectural decisions from every past meeting and cross-references them against new transcripts using technology mutex groups — flagging conflicts like PostgreSQL versus MongoDB or JWT versus session auth before they cause problems. EchoSync supports Google Gemini, Groq, IBM Watsonx Granite, and Ollama. Watsonx powers meeting sentiment scoring and per-task developer context generation. Every GitHub issue includes AI reasoning, story points, a branch name, a commit template, and the exact transcript quote that created it.
17 May 2026
